See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/337372992

An In Vitro Comparison of Estrogenic Equivalents Per Serving Size of Some

Common Foods

Article  in  Journal of Food Science · November 2019

DOI: 10.1111/1750-3841.14847

CITATIONS READS

2 136

4 authors, including:

Nancy W Shappell Eric P Berg

United States Department of Agriculture North Dakota State University
51 PUBLICATIONS   1,255 CITATIONS    120 PUBLICATIONS   1,845 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Risk Assessment of drug residues in milk and milk products in the United States View project

All content following this page was uploaded by Eric P Berg on 26 January 2021.

The user has requested enhancement of the downloaded file.

 An In Vitro Comparison of Estrogenic Equivalents
Per Serving Size of Some Common Foods
Nancy W. Shappell , Eric P. Berg , James D. Magolski, and Lloyd O. Billey

Abstract: The public assumes that some foods, such as milk and ground beef from cattle receiving steroidal implants,
are associated with estrogenic hormones, while other foods are presumed “safe” or nonestrogenic. Here, we investigate
these assumptions by assessing the relative estrogenic activity of a serving size of four foods: skim milk (8 oz), rice (48 g
dry wt) in cooking bag, ground beef patties from steers raised with or without hormone implantation (quarter lb each,
114 g), and tofu burgers (isocaloric to beef burger, 198 g), using an in vitro assay (E-Screen). Mean picogram (pg) estradiol
equivalents (E2Eqs) on a serving basis were as follows: skim milk 120; rice 400; rice prepared in cooking bag 370; rice
boiling bag alone 4 pg per bag, ground beef burger (obtained from the tissue of cattle that had received no hormone
implants) 389, beef burger (obtained from cattle that had received hormone implant) 384, and tofu burger 1,020,000.
Rice E2Eqs were highly variable, but the plastic cooking bags provided by the manufacturer added negligible E2Eqs. The
source of estrogenic activity in rice may have been due to contamination with the mycotoxin zearalenone. The E-Screen
E2Eqs of tofu burger extracts agreed with those predicted based on chemical concentrations of the most estrogenic
component times their E2Eq factor. While a tofu burger contained around three times the estrogenic activity of a daily
dose of estrogen replacement therapy (125 mg, Premarin R
, 303,000 pg); the other foods––a quarter pound ground beef
burger at approximately equal calorie count, a serving of milk, or rice, were all at least 750-fold less estrogenic.

Keywords: beef, E-Screen, milk, phytoestrogens, rice, tofu

Practical Application: When consuming the recognized serving size of a food, how much estrogenic activity can we
expect? While the public assumes that some foods, such as milk and ground beef from cattle receiving steroidal implants,
are associated with estrogenic hormones, other foods are presumed “safe” or nonestrogenic. Using one assay, a tofu burger
contained three times the estrogenic activity of a dose of hormone replacement therapy commonly prescribed for women
after hysterectomy or menopause (Premarin R
); while other foods––a quarter pound ground beef burger at approximately
equal calorie count, a serving of milk, or rice, were all at least 750-fold less estrogenic.

Introduction
For more than a decade, scientists have been concerned with
consumer exposures to endocrine active substances in foods. As
a result, much data have been published documenting concentra-
tions of such chemicals in foods, including investigations of the
biological effects of the pure chemicals. While in the past, con-
sumption of naturally occurring hormones in food was typically
regarded as innocuous, the introduction of synthetic hormones in
livestock production to promote growth beginning in the 1950s
drew concern––focusing on diethylstilbesterol (DES), a synthetic
estrogen. DES use was found to increase average daily gain of
both steers and heifers (Raun & Preston, 2002). Unexpectedly,
the uterotropic effect of feeding tissue of DES-treated animals to
ovarectomized mice (an increase in uterine weight) was found to
be equivalent to feeding an all grain diet (Stob, Andrews, Zarrow,
JFDS-2019-1034 Submitted 6/28/2019, Accepted 9/29/2019. Authors Shappell
(retired) and Billey are with USDA, Agricultural Research Service, Edward T. Schafer
Agricultural Research Center Biosciences Research Laboratory, 1616 Albrecht Boule-
vard, Fargo, ND 58102, USA. Author Berg is with Dept. of Animal Sciences, North
Dakota State Univ, Fargo, ND 58108, USA. Author Magolski was with Dept. of
Animal Sciences, North Dakota State Univ., Fargo, ND 58108, USA. and now he
is presently with Coleman Natural Foods, Westminster, CO 80234, USA. Direct
inquiries to author Shappell (E-mail: nwshappell@gmail.com).
No outside funding was received for this project.
Toxicology & Chemical
Food Safety
& Beeson, 1954). Physicans prescribed DES to women with a his-
tory of miscarriage to aid in the maintenance of pregnancy (CDC,
2012). A risk to human health was implicated as the U.S. Natl.
Cancer Inst. cited DES to be associated with increased risk of can-
cers in the daughters of women who took DES while pregnant
(National Cancer Institute, 2011).
Concerns about hormones in food have increased in recent
years especially related to the use of natural and synthetic steroids
(mimicking the action of the natural hormones testosterone,
17-β estradiol [E2], and progesterone) in cattle. The Food and
Drug Administration (FDA) has approved these treatments based
on documentation that tissue concentrations at time of slaugh-
ter are of no human health consequence. However, consumer
fears related to early puberty and cancer have been raised over
this practice. Stephany (2010) determined the dose of the nat-
ural hormone E2 to be 2.5 ng from a 250 g steak from cattle
that had not received exogenous hormones, compared to 5 ng
from 250 g from hormone-treated cattle. The same study found
a 50 g hen’s egg contains 6.5 ng of E2. These data indicated
that the consumption of a natural egg might be expected to con-
tribute more E2 than beef, whether from treated or untreated
cattle. Because of concern over hormone treatment of cattle,
some consumers have turned to meat substitutes, such as tofu,
long ago documented (Murphy et al., 1999) to contain high con-
centrations of phytoestrogens (components of plants that are estro-
genic). The present study uses an in vitro assay, the E-Screen (Soto

Published 2019. This article is a U.S. Government work and is in the public domain in the USA.
3876 Journal of Food Science r Vol. 84, Iss. 12, 2019 doi: 10.1111/1750-3841.14847
Further reproduction without permission is prohibited
Estrogenic equivalents per food serving . . .
et al., 1995), to measure estrogen-dependent cellular proliferation,
an integrated cumulative cellular response, rather than other assays
for estrogen activity that quantitate estrogen receptor binding, or
by transcriptional activation via an estrogen receptor, typically in
transfected cell lines. The E-Screen activity of a small sample of se-
lected foods will be evaluated on a human nontransfected cell line.
The rationale for food choices were as follows. Cow milk was
chosen because variable results have been published on both its es-
trogen content and bioactivity (Capriotti et al., 2015, Chen et al.,
2014, Furnari, Maroun, Gyawali, Snyder, & Davis, 2012, Goyon
et al., 2016, Stypula-Trebas, Minta, Radko, & Zmudzki, 2015),
yet milk has long been a staple of human diets. Because consumer
concern has risen over the use of steroid hormone implants in
beef cattle, reflected by internet postings, such as the one by Blair
(2015), South Dakota State Univ. Extension entitled “Hormones
in Beef- myth vs fact”, ground beef burgers from steers raised
with or without steroid hormone implants were evaluated. For
comparison to the ground beef, tofu, the popular meat substitute,
was prepared as burgers and assayed. Finally, public concern
over the contribution of plastics to the estrogenic contamination
of foods (via leaching of chemicals, for example, bisphenol A
and phthalates) led us to evaluate the estrogenic contribution of
boiling bags supplied with some rice products. Rice was prepared
either with or without using the “boiling bag” to ascertain
if the bag contributes measurably to the estrogenic activity of
rice.
For a frame of reference, the estradiol equivalent (E2Eq) of one
“dose” of Premarin R
, a hormone replacement therapy (HRT)
often prescribed to menopausal women, was measured. We have
also examined how well-predictive estimates of estrogenic activity,
based on estrogenic chemical concentration data (either measured
or reported in the literature), compare with extract estrogenic ac-
tivity measured by E-Screen. Finally, a crude estimate of potential
circulating E2Eq concentrations has been calculated for compar-
ison to the range of normal estradiol concentrations in children,
men, and women.
Materials and Methods
Extractions
Milk. Pasteurized, homogenized skim milk from one Midwest-
ern United States distributor was purchased from grocery stores
on five different dates (mean 18 ± 2.0 days prior to last sell by
date). Milk was labeled as coming from cows not treated with
“rBST/rBGH” (recombinant bovine somatotropin or recombi-
nant bovine growth hormone). Milk (125 mL) was diluted with
an equal volume of nanopure water (npH2O), mixed, and then
extracted and subsequently concentrated using solid-phase extrac-
tion (SPE, OASIS HLB cartridge, 200 mg, 5 cc, glass, Waters,
Milford, MA, USA). Columns were preconditioned by sequen-
tial application of 6 mL each of: acetone, methanol, acetonitrile,
ethyl acetate, methylene chloride, and tert-butyl methyl ether fol-
lowed by 6 mL of npH2O. After the sample was applied to the
column, the eluate (3 mL each of the same solvent series, applied
sequentially) was collected and evaporated to dryness under N2
at 37 °C. Dry residues were stored at –20 °C, until the time of
E-Screen analysis, when they were resuspended in 160 µL sterile
npH2O.
Tofu. Four brands of tofu were purchased at a retail outlet.
Patties (198 g precooking weight) were cooked in stainless steel
pans to an internal temperature of 76 °C, and dry matter was
determined on quintuplicate samples of each cooked patty (com-
positional analysis, Table S1). The extraction protocol was pre-
viously described for isoflavone analysis of foods (Murphy et al.,
1999) and modified by an additional hexane extraction step. A 3 g
(postcooking weight) portion of crumbled patty was extracted as
described in the Supporting Information. The aqueous fraction
was diluted to approximately 90 mL and concentrated by SPE and
evaporated to dryness as described for milk. Dry residues were
stored at –20 °C and resuspended at the time of E-Screen analysis
in 380 µL npH2O and 20 µL dimethyl sulfoxide (DMSO, Sigma
Chemicals, St. Louis, MO, USA) to enhance the solubility. Extrac-
tion efficiency and/or matrix interference in the E-Screen assay
was assessed on separate samples that were fortified with synthetic
genistein (800 µg/3 g of cooked patty); stock solution 2.5 mg/mL
(99.1% purity by HPLC, Sigma Chemicals). Genistein fortifica-
tion was sufficient to ensure that estrogenic activity in fortified tofu
was greater than the inherent estrogenic activity from the genis-
tein already present in the tofu samples. Initially, methodology was
tested on one patty (Brand D) from which triplicate extractions
were performed with and without genistein fortification. Subse-
quently, three separate patties for each of three brands A to C
were prepared and extracted with and without fortification. Ex-
tracts were also analyzed by liquid chromatography tandem mass
spectrometry (LC-MS-MS) for quantitation of the phytoestrogens:
genistin and genistein (method details are given in Table S2).
Ground beef. Initial ground beef samples had been prepared
for a feeding trial (Magolski et al., 2014) from the deboned car-
casses of steers (2 ea) raised either with or without the use of
implants containing an anabolic steroid and estrogen (Synovex
ChoiceTM , 100 mg of trenbolone acetate and 14 mg of estradiol
benzoate, Fort Dodge Animal Health, Fort Dodge, IA, USA). Im-
plants were administered in standard approved manner: at weaning
(approximately 230 kg body weight) and again at 410 kg of body
weight. Details about all cattle are given in Table S3. Subsequent
ground beef samples were prepared from individual organically
reared steers (nonimplanted, n = 7), and individual steers that had
received anabolic implants (n = 4, implanted once at weaning
with Synovex One FeedlotTM , containing 200 mg of trenbolone
and 28 mg of estradiol benzoate). Ground beef patties from indi-
vidual steers were prepared in triplicate, while a single patty was
prepared from composite samples. Beef patties were pan-cooked
as described for tofu patties and dry matter was determined in
quintuplicate on cooked samples (Table S4). Cooked burgers were
extracted in a modification of a method for extraction of estrogens
from fish (Al-Ansari, Saleem, Kimpe, Trudeau, & Blias, 2011). A
portion of the crumbled patty (5 g postcooking weight) was mixed
with approximately 2 g of solvent-extracted diatomaceous earth
(DE extracted on an Accelerated Solvent Extractor 200 [ASE],
Dionex Corp., Salt Lake City, UT, USA). The DE/sample mix-
ture was extracted as described in the Supporting Information.
The aqueous material from the ASE extraction was diluted to
100 mL with npH2O and processed by SPE as described for
milk. Evaporated eluate residues were stored at –20 °C and re-
suspended in 400 µL of npH2O on the day of E-Screen analysis.
Extraction efficiency and/or matrix interference was assessed by
fortifying ground beef with deuterated E2 (17d4 β-E2 , CDN Iso-
topes, Pointe-Claire, Quebec, CA; 600 pg/5 g of beef postcooking
weight) from both implanted and nonimplanted cattle. On orig-
inal composite samples, three extractions were performed on the
nonimplanted beef and six extractions were performed on im-
planted beef, as the latter exhibited higher variability in E2Eqs.
For the subsequent individual steer samples, triplicate burgers were
Toxicology & Chemical
typically extracted only once.
Food Safety
Vol. 84, Iss. 12, 2019 r Journal of Food Science 3877
 Estrogenic equivalents per food serving . . .
Rice. Six boxes of brand name Boil-in-Bag-style rice were
purchased from different retail outlets in the Midwestern United
States. Though three of the six boxes were from the same lot,
each box was considered a unique sample, as it reflected what one
household would consume. Duplicate sample sets were prepared
for each box consisting of rice prepared inside the bag (accord-
ing to the manufacturer’s microwave directions), without the bag,
or bag alone. Cooked rice was extracted using the same series
of solvents used for the OASIS HLB SPE of other foods as fol-
lows. One serving was weighed (approximately 110 g cooked rice
or approximately 1 cup) and extracted with 200 mL of acetone
for 15 min (shaking at room temperature, setting 5 in a Dubnoff
Metabolic Shaking Incubator, Precision Scientific, Chicago, IL,
USA) and then decanted into a 500 mL round bottom flask. Ex-
tractions were repeated using 100 mL each of the following solvent
series: methanol, acetonitrile, ethyl acetate, and methylene chlo-
ride, using a clean flask for each solvent. Solvents were volatilized
sequentially in a common flask using a rotavapor (RE 111, Büchi,
Flawil, Switzerland) at 40 °C and the final aqueous volume was
recorded. The cooking bags were emptied of their contents and
bags were microwaved in 946 mL of npH2O in the same manner
as rice in bags, to determine any potential contribution of the bag
to total estrogenic activity of the rice. Once the water was cooled,
it was processed by SPE as described for milk. Evaporated residues
were stored at –20 °C and resuspended in 400 µL of npH2O at
the time of E-Screen analysis. Extracts were later analyzed by LC-
MS-MS for zearalenone and zearalenol (method details are given
in Table S2).
Conjugated estrogens (Premarin R
). The pharmaceutical
used for HRT in postmenopausal women (a formulation of con-
jugated estrogens in tablet form) was evaluated by E-Screen. The
product, Premarin R
, is labeled 1.25 mg of ingredients in a pro-
prietary mixture formulated to a consistent, specific, biological
activity. One dose (pill) was mechanically pulverized and dissolved
in 10 mL of sterile npH2O and shaken until dissolved. Solutions
were held overnight at 4 °C and diluted in media immediately
prior to cell treatment. Multiple dilutions were tested, but lin-
ear cellular proliferation to the conjugated estrogens was obtained
from a 1:15,000 to 1:30,000 dilution of the original 10 mL so-
lution containing one pill. Four individual pills of Premarin were
pulverized, diluted, and tested separately over multiple E-Screen
experiments.
E-Screen
The MCF-7 BOS, estrogen-dependent cell line (human mam-
mary epithelial cell line, graciously provided by Drs. Ana Soto
and Carlos Sonnenschein, Tufts Univ., Boston, MA, USA), was
used to determine estrogenicity relative to E2 as previously de-
scribed (Shappell, 2006). Briefly, resuspended extracts were di-
luted in cell culture media (Dulbecco’s Modified Eagle Medium).
Twenty-four hours post-plating, steroid-containing medium (fetal
bovine serum, FBS) was removed from cells and replaced with
steroid-free medium (minus phenol red and containing charcoal
dextran-stripped FBS) containing diluted extracts of samples or
17β-E2 (standard curve from 1 × 10−13 to 1 × 10−9 M). Af-
ter 5 days of incubation, cells were fixed with trichloroacetic acid,
stained for protein with sulforhodamine B (Sigma Chemicals), sol-
ubilized in buffer, and absorbance measured (490 nm). All plates
contained control wells with cells receiving only stripped serum to
confirm a lack of proliferative response in the absence of estrogenic
stimuli. E2Eqs were determined based on a regression analysis of
Toxicology & Chemical
the 17β-E2 curve from the same experiment. Extracts were tested
Food Safety
3878 Journal of Food Science r Vol. 84, Iss. 12, 2019
over a wide range of dilutions, and those resulting in absorbance
readings in the linear range (1 × 10−12 to 1 × 10−11 M) were used
for interpolation of the data. Evidence of toxicity was determined
by evaluating cellular proliferation in the presence of the sam-
ple spiked with 17β-E2 versus 17β-E2 alone. Estrogenic activity
was calculated on dilutions not exhibiting toxicity. The estrogen-
dependence of cellular proliferation was confirmed in all cases by
coincubation with the E2-receptor antagonist ICI 182,780(Tocris,
Ellisville, MO, USA) (Rassmussen & Nielsen, 2002) and found to
ablate all sample-dependent proliferative activity. Milk and rice
contained additives (vitamin A and D, folic acid, niacin, and thi-
amine) that were evaluated separately by E-Screen at concentra-
tions equivalent to those found “as consumed,” as well as at the
concentrations estimated to be present in milk and rice as assayed
(details are given in the Supporting Information). Relative E2Eqs
of compounds was previously published (Shappell, Mostrom, &
Lenneman, 2012) with the exception, trenbolone. Representative
dose–response curves are presented in Figure S1.
Results and Discussion
Recognizing that the samples used in this study cannot be
presumed to represent these foods across all regions, countries,
production facilities, and so on, results will be discussed in com-
parisons to available literature values for the various estrogenic
biochemicals present in these foods.
Milk. The E2Eqs of milk had an average concentration of
520 ± 200 SD pg/L (Table S4), ranging from 54 to 184 pg/8 oz.
serving (236 mL) in extracts from five different dates (Figure 1A).
The coefficient of variation (COV) of milk E2Eqs from five dif-
ferent purchase dates was 39%, and within assay variability for a
sample was 15%. Vitamin A and D supplements were not estro-
genic, either at dilutions used to quantitate E2Eqs in milk extracts
or at the supplemented concentrations present in commercially
sold milk. In the United States, commercial milk is transported in
bulk tanks and dairy herds are maintained to provide production
throughout the year. Therefore, a composite of milk from cows at
various stages of lactation and pregnancy will be present within the
bulk tank. Depending on the state of pregnancy, E2 concentra-
tion of homogenized whole cow’s milk was reported ranging from
nondetectable to 22.3 ng/L (SPE extraction followed by radioim-
munoassay) (Pape-Zambito, Magliaro, & Kensinger, 2007). Across
stages of pregnancy, values ranged from 0.5 ng/L (1st trimester)
to 5.0 ng/L (3rd trimester) averaging 1.4 ng/L (Pape-Zambito,
Magliaro, & Kensinger, 2008). The concentrations of E2Eqs in
the skim milk samples in the present study (mean 0.52 ng/L) were
similar to those reported for 1st trimester cows. The most re-
cent literature remains inconsistent as to the E2 concentration of
cow milk (Capriotti et al., 2015, Chen et al., 2014, Goyon et al.,
2016). Whole milk analyses of samples from across 48 United
States reported E2 ranged from 5.0 to 6.4 ng/L (Vicini et al.,
2008), while others reported skim milk concentrations as 55%
lower (Wolford & Argoudelis, 1979). Using the lowest and high-
est reported values, 5.0 and 22 ng/L, and multiplying those values
by 16% (the percentage of unconjugated E2 Courant et al., 2007)
and then times 55% to adjust for difference in concentration of E2
in skim instead of whole milk, one could expect between 0.4 and
2.0 ng/L of 17β-E2 or E2Eqs in skim milk. Again, our findings
by E-Screen (0.52 ng/L E2Eqs) fall within this range. It is not sur-
prising that Stypula-Trebas et al. (2015) failed to detect estrogen
activity in raw, tank, pasteurized, or ultrahigh temperature milk
samples ± fortified with 1 ng/mL of E2 using the in vitro Yeast
Estrogen Screen assay, as the assay is much less sensitive, with a
Estrogenic equivalents per food serving . . .

Figure 1–E-Screen determination of estrogenic activity of foods on a serving basis, mean pgs for all except Tofu, that are ng ± S.D. E2Eqs of all extracts
were calculated from concurrently run 17β-estradiol (E2 ) standard curve. (A) Skim milk 8 ounces. (B) Tofu burgers 144 g of cooked burger (A to C
represent extracts of three burgers, while D∗ represents extract of one burger). (C) Ground beef burgers prepared from steers raised with (IMP) or
without (NON) growth-promoting implants (SynovexTM ). The composite samples were prepared from two nonimplanted steers (A) and two implanted
steers (B) E2Eq of mean of multiple extractions from one available burger. Bars B1 to 4 and C1 to 3 represent the mean E2Eqs from three separate
burgers that were each prepared from individual nonimplanted organic steer, and D1 to 4 were from implanted steers all of the same age and harvested
at the same time. E2Eqs pg per burgers (precooked wt 114 g). (D) Boil-in-Bag Rice 48 g dry rice, bars with the same letter were from the same lot.
Estrogenic activity below the limits of quantitation or detection is represented on the graph as “–20 pg”/serving. Estrogenic activity of bag alone was
below the limits of quantitation, estimated as 4.1 ± 2.27 pg/serving.

limit of detection and EC50 reported as 270 ng/L. MCF-7 cells
were previously used to assess estrogenic activity of milk in two
papers. One group detected no activity (Tate, Bibb, & Larcom,
2011), possibly due to shortened incubation period (only 48 hr
exposure), though E2 was not evaluated as a positive control, and
therefore laboratory E2 EC50 (concentration when E2 induced
½ of maximal proliferation) could not be compared. The second
paper reported findings in agreement with ours, median concen-
tration of 0.55 ng/L by E-Screen, n = 3 (Courant et al., 2007).
Tofu. Though phytoestrogen content of tofu has been well
documented (Murphy et al., 1999), this is thought to be the
first report of tofu’s estrogenic activity using an in vitro assay. The
precooked weight of the tofu burger was approximately 198 g,
with 144 g postcooking. The mean E2Eq of a tofu burger was
7,080 pg/g wet wt, a dose of E2Eqs translating to 1,020,000 ±
259,000 pg if consumed in the typical isocaloric serving size
represented by a quarter-pound ground beef burger (data by
brands Figure 1B, overall mean Figure 2, and Table S4). Moisture
content among different tofu brands was variable. Cooked patties
of three brands had approximately 20% dry matter, while the
fourth brand had only approximately 14% dry matter. Ironically,
this brand was found to have the highest estrogenicity on an
either wet or dry basis.
The two most estrogenic phytoestrogens in tofu, genistein and
glycosylated genistein, genistin, were measured at 4 ± 4.0 and
84 ± 12.6 µg/g wet wt, respectively, in tofu burger extracts (LC-
MS-MS, Table S5). The higher variation measured in the genis-
tein content was due to Brand D, which had approximately four
times greater genistein than the other brands. Good agreement was
found between the cooked patty concentrations of genistein and
genistin and those reported by Murphy et al. (1999) for uncooked
tofu, 11.2 ± 7.9 and 102 ± 21.1 µg/g wet wt, respectively.
In order to assess how well the biochemical measurement of
genistein and genistin content of tofu could predict E-Screen es-
trogenicity, previously reported phytoestrogen concentrations in
tofu and those from this study of cooked tofu were multiplied by
their E-Screen determined E2Eq factors. The factors for genistein
and genistin (found to be deconjugated to genistein by FBS present
in the E-Screen assay medium) are 1.5 × 10−4 and 1.54 × 10−4
E2Eqs, respectively (Shappell et al., 2012). Summing the estro-
genicity expected from genistein and genistin concentrations as
measured by LC-MS-MS, the cooked tofu burger theoretical
E2Eqs were 13.5 ng/g cooked burger wt versus 17.4 ng/g E2Eqs
using the literature for genistein and genistin uncooked tofu con-
centrations (Murphy et al., 1999). The cooked tofu E2Eq estimate
is 78% of the literature value for uncooked tofu, commensurate
with the 73% of starting weight after cooking (198 to 144 g cooked
wt). Previously, E2Eqs by E-screen were found to accurately reflect
estimates based on phytoestrogen content, with an R2 value of
0.91 (Shappell et al., 2012). Extraction recoveries, assessed by
Toxicology & Chemical
Food Safety

Vol. 84, Iss. 12, 2019 r Journal of Food Science 3879

 Estrogenic equivalents per food serving . . .

Figure 2–Estrogenic activity of foods on a wet weight (A) and serving size basis (B) and Premarin R
. Mean estrogenic activity is represented with S.D.
value inside of bars, with median values above the bar. Serving size: skim milk (8 ounces); cooked tofu patty (198 g precooked patties); a “quarter-pound”
hamburger from nonimplanted steers (NON) or from steers that had growth-promoting implants (IMP) (114 g precooked weight); white rice (1/4 cup
or 48 g dry rice) prepared with or without commercial “boil in bags” and bags alone; and one “1.25 mg” daily dose of the hormone replacement
pharmaceutical Premarin R
. “Safe dose” based on Joint FAO/WHO Expert Committee on Food Additives (2000b).

fortification with genistein, were 36% chemically and 62% by
E-Screen, respectively (Table S4). The genistein used for sam-
ple fortification was contaminated with genistin (determined by
LC-MS-MS, data not shown). The contaminating genistin in the
fortified sample would be converted to genistein in the presence
of FBS, leading to a higher proliferative effect on the cells and
therefore a higher “recovery of estrogenic activity” than recovery
based on chemical measurement of fortified genistein.
The serving size for this study was chosen to approximate
isocaloric content of a quarter pound ground beef burger, but
it is larger than the serving size listed by the manufacturers (85 g).
Assuming similar weight loss with cooking, the label serving size
cooked would be approximately 62 g, with an expected approx-
imately 438,000 pg of E2Eqs. While the tofu burger “serving”
was approximately 2.5 times the label serving size, one brand of
Toxicology & Chemical
ported health benefits of soy protein. The health benefits or risks
of tofu consumption have yet to be irrefutably established (Zaheer
& Akhtar, 2017) and are not a topic of consideration in this paper.
Ground beef from nonimplanted steers. Ground beef
hamburger (114 g uncooked and 71 g cooked) contained 5.5 ±
2.4 pg/g wet wt or 389 pg of E2Eqs per serving (Figure 1C, overall
mean Figure 2, and Table S4). Literature values for E2, estriol, and
estrone in raw beef have been reported as less than the detection
limits of 100 pg/g wet weight (Seo, Kim, Chung, & Hong, 2005)
or ࣘ10 pg/g (Hartmannn, Lacorn, & Steinhart, 1998). Though
cooking was not found to degrade estrone or 17β-E2 (fortified
at 100 ng/g in raw ground beef), when cooked to an internal
temperature of 72 °C, estrogen recovery was reduced to 70% –
75% of fortification values, which correlated with the reduction
in patty weight due to loss of fat and juices with cooking (analyses

tofu recommended consumption of 2.5 servings to obtain the pur- by LC-MS-MS) (Braekevelt, Lau, Tague, Popovic, & Tittlemier,
Food Safety

3880 Journal of Food Science r Vol. 84, Iss. 12, 2019

Estrogenic equivalents per food serving . . .
2011). Therefore, the concentration would remain equivalent in the
cooked and uncooked patty and values for uncooked ground beef
(ࣘ10 pg 17β-E2 /g (Hartmannn et al., 1998) should be comparable
to the mean E2Eq determined here. This value, even when cor-
rected for extraction losses and/or matrix effects (8.6 pg E2Eqs/g
based on 57% recovery value, Table S4), is well below the U.S.
tolerance limit of 17-β-E2 in raw beef of muscle and fat (120 and
480 pg/g, respectively) (Code of Federal Regulations, 2011).
Ground beef from implanted steers. Mean E2Eqs of ex-
tracts of ground beef prepared from steers receiving the growth-
promoting hormones (5.7 ± 4.0 pg/g wet wt) were similar to that
of nonimplanted ground beef (5.5). The composite beef sample
exhibited higher variability in multiple extractions, with the high-
est implanted ground beef E2Eqs of 28 pg/g wet wt and median
E2Eq of 12.8 pg/g wt wet, while the nonimplanted median was
8.4 pg/g wet wt (Figure 1C, on a serving basis). The range of
all extracts from implanted cattle was 3 to 28 pg E2Eq/g wet wt
(Table S4). COV of E-Screen assay of an extract (within a plate)
was similar for beef from both nonimplanted and implanted cattle
(10% and 11%, respectively), but the COV of extracts from the
nonimplanted cattle was higher (COV 34%) than from implanted
cattle (18%). The variability could be a result of different amounts
of blood entrained in the patties, or might reflect more homo-
geneity of tissue concentrations in animals that are implanted. The
theory of implanted cattle having less variable E2Eqs was sup-
ported by the data from a burger of an implanted heifer. Due to a
communication oversight, one ground beef sample had been ob-
tained from a heifer rather a steer. Commercially prepared ground
beef will include trimmings from cattle representing a wide range
of age, gender (male, female, and castrate), and production envi-
ronment. The E2Eqs of the ground beef from the heifer (2.5 ±
0.7 pg/g wet wt of triplicate burgers) were in line with the average
of the “companion” steers (raised at the same time, same facility,
3.9 ± 0.6). Statistical analysis is not required to see there was no
difference in E2Eq of ground beef from implanted versus nonim-
planted steers. The overall mean weight of implanted beef patties
was 70 g, resulting in 384 pg E2Eqs per patty serving (Figure 2).
In documentation provided to the U.S. FDA for the approval
process for implant product Synovex R
Plus (1996), reporting tis-
sue concentrations, steers received implants with the same steroids
used in this study, at 1.5 times the marketed concentration. This
equates to three times the concentration of a single implant used
for the composite steers and 1.5 times that of the second set of
steers. They reported E2 concentration in muscles of 24 pg/g
and in fat of 95 pg/g when slaughtered 30 days postimplantation,
which, when adjusted for dose differences, would yield theoret-
ical E2Eqs concentration per g of uncooked implanted beef of
8 to 18 pg/g, similar to the range in our study of 3 to 28 pg/g
measured.
The SynovexTM implants also contain trenbolone acetate to de-
crease the percentage of carcass fat (Bryant et al., 2010) and may
be reflected in the reduced fat content of the cooked ground
beef (11.2% compared with 13.9% in nonimplanted beef, Ta-
ble S1). Fortified estrogen loss in cooking has been correlated to
fat content, with only approximately 5% loss found in 95% lean
ground beef and 25% to 30% loss in 75% lean beef (Braekevelt
et al., 2011). Regardless of the fat content, in this study, the
E2Eq per serving was equivalent, whether the patty was prepared
from an implanted animal or nonimplanted animal (384 compared
with 379 pg E2Eqs, respectively). The highest concentration of
27.8 pg/g in the cooked patty if adjusted for extraction losses
and/or matrix effects (30% recovery, Table S4) would be 93 pg/g
cooked burger. These values are approximately 20 to 500 times
less than the EU Natl. Minimum Required Performance Limits
of 2 to 50 ng/g for xenohormones in raw products and below the
U.S. residue limit for estrogen (Impens, De Wasch, Cornelis, &
De Brabander, 2002). The allowable incremental increase of E2
over concentrations present in tissues of untreated animals was es-
tablished by the United States under CFR 556.240 (Code of Fed-
eral Regulations, 2011) as not to exceed 120 pg/g for muscle or
480 pg/g for fat in uncooked meat from cattle. Because E2Eqs
from patties from either implanted or nonimplanted animals were
essentially equal, on either a mean or median basis, the conditions
of Code of Federal Regulations (2011) were met.
Though the concentration of trenbolone acetate in the implants
used in cattle for this study is about seven times that of E2 (100 mg
of trenbolone acetate and 14 mg 17β-E2), trenbolone has approxi-
mately 1/50,000 the activity of E2 on a molar basis in the E-Screen
assay (data not shown, trenbolone assayed by E-Screen from 1 pM
to 10 µM). Trenbolone concentrations were reported as 2 and
1.4 ng/g in muscle and fat, respectively, in beef from implanted
steers (Synovex Plus, 1996), equating to a theoretical 0.04 pg/g
E2Eqs. Based on these calculations, increased estrogenic activity
from trenbolone residues would be insignificant relative to that
contributed by naturally produced estrogen.
Rice. The estrogenic activity of the commercial rice analyzed
was highly variable; even within a package. The actual concen-
tration of E2Eq for rice prepared by the bag method, or without
the bag ranged from 0 to 66 and 0 to 8 pg/g wet wt, respectively
(Table S4), equating to 0 up to 7,303 pg of E2Eqs per 110 g wet wt
serving (Figure 1D). Median values per serving were 104 and 400
pg, prepared with or without the bag, respectively. None of the
rice supplements (folic acid, niacin, and thiamine) contributed es-
trogenic activity (details are given in the Supporting Information).
Previously, no data were available on the estrogenic activity con-
tributed by cooking bags to the overall estrogenic burden of rice.
E-Screen assay of water from bags cooked without rice did con-
tain quantifiable estrogenic activity (limit of quantitation = 2 pg
bag), though at worst, a bag was found to contribute no more
than 4 pg of E2Eqs per serving of rice (Figure 1D). While
there may have been a tendency for lower estrogenic activity of
rice prepared in the bag, the highly variable results from sam-
ple to sample (within box) would negate any relevance of such
findings.
Because rice does not contain known phytoestrogens, it was
thought that estrogenic activity measured in the rice samples
might have originated from mycotoxins. Using LC-MS, the estro-
genic mycotoxin zearalenone was detected in some rice extracts.
Zearalenone concentrations were below the limits of quantitation
(0.2 ng/g) in all but three of the six extracts from one lot of rice
(three prepared with and three without bag). Concentrations of
these extracts ranged from 0.23 to 0.7 ng/g rice, or an estimated
11 to 35 ng zearalenone per serving of rice. Two of the three
correct mass fragments used to identify zearalenone were present
in the rice only extracts from two of the other three lots. Us-
ing the E2Eq for zearalenone determined by our laboratory of
0.01 (Shappell et al., 2012), zearalenone concentrations were not
high enough to account for the activity measured by E-Screen.
In a review of zearalenone contamination of various feedstuffs
around the globe, rice was rarely found to be contaminated, in
contrast to other foods, such as corn, and when detected in rice
(Joint FAO/WHO Expert Committee on Food Additives, 2000b),
concentrations ranged from 15 to an extreme of 3,060 µg/kg (one
Toxicology & Chemical
sample only).
Food Safety
Vol. 84, Iss. 12, 2019 r Journal of Food Science 3881
 Estrogenic equivalents per food serving . . .
Table 1–Theoretical E2Eqs in blood (pg/mL) if 10%a of E2Eqs absorbed.
Womanb
circulating E2 range:
15–350 pg/mLd
Food E2Eq
(serving size) E2Eqs/serving (pg/mL)
Tofu cooked (precooked wt = 1 serving, 12
85 g) 438,000 pg
Tofu cooked burger (precooked wt = 2.3 28
servingsf , 198 g) 1,044,000 pg
Ground beef burger – NON(114 g 0.010
precooked wt) 389 pg
Ground beef burger – IMPg (114 g 0.010
precooked wt) 384 pg
Skim milk (236 mL, 8 oz.) 124 pg 0.003
Rice (48 g dry or ¼ cup dry) 1,439 pg 0.04
a
Conservative estimate, especially for absorption of phytoestrogens.
b
59 kg, 165 cm, blood volume, 3,740 mL (Blood Volume Calculator).
c
82 kg, 178 cm, blood volume 5,480 mL (Blood Volume Calculator).
d
E2 values, blood estradiol values (Mayo Clinic).
e
30 kg, 109 cm, blood volume, 1,640 ♀ or 2,040 ♂ mL (Blood Volume Calculator).
f
Based on label serving size of 8.5g.
g
Ground beef was prepared from unimplanted steers (NON) or implanted steers (IMP).
Foods versus Premarin R
and their modeled circulating
E2Eqs versus circulating E2 concentrations in humans.
To provide meaning for the relative estrogenic activity of various
foods, it is helpful to compare them to one another. For context,
Figure 2 shows the estrogenic activity of each food on a serving
basis and the estrogenic activity of one “dose” of Premarin R
, the
HRT used for postmenopausal women. A 2001 report by the Natl.
Center for Health Statistics (U.S. Center for Disease Control, Brett
& Chong, 2001) cited that 40% of postmenopausal women in the
United States between 1988 and1994 had used oral HRT. The
same report states “Until we have more definitive information
about HRT’s effect on diseases, each woman must, with the help
of her physician, weigh the risks and benefits of use.” Previously,
steroid intake from a variety of foodstuffs was compared to the daily
production of the hormones in a review (Hartmannn et al., 1998)
in which Kushinsky (1983) reported that estrogen production
(E2 + estrone) in prepubertal children and men ranged from 54
to 140 µg/day up to 630 µg/day for women. Daily intake of the
estrogens ranged from 0.07 to 0.10 µg/day (or a maximum 0.2%
of production). If three servings of skim milk and one serving of
each of the other foods presently assayed were consumed, using the
highest measured E2Eqs, the total would be 1.034 µg E2Eqs/day,
approximately 10 times the 0.1 µg estimated by Hartmannn et al.
(1998). But, of the total estrogenic activity, the tofu burger ac-
counted for >98%. Without the tofu burger, the sum of E2Eqs
would be only 0.014 µg, or approximately 10% of the literature
value for estimated estradiol consumption, and less than 5% of the
E2Eqs from one Premarin R
tablet. The Joint FAO/WHO Expert
Committee on Food Additives established an average daily intake
for E2 of 0 to 50 ng/kg body weight per day based on a No
Observable Effect Level of 0.3 mg/day and a safety factor of 10
for sensitive populations (Joint FAO/WHO Expert Committee
on Food Additives, 2000b). The 1.034 E2Eqs µg/day from con-
sumption of the full complement of foods tested (including three
servings of milk) would equate to 3.4% of the 0.030 mg/day “safe
dose” for 17β-E2.
Further context for comparison is provided by considering con-
centrations of the synthetic estrogen prescribed to disrupt the re-
productive cycle of women, ethinylestradiol (EE2), a component
Toxicology & Chemical
Food Safety
3882 Journal of Food Science r Vol. 84, Iss. 12, 2019
Manc Childe
circulating E2 range: circulating E2 range:
10 to 40 pg/mLd ♀, ♂ ࣘ 20 pg/mLd
E2Eq E2Eq
(pg/mL) (pg/mL)
8 9, 8
19 21, 17
0.007 0.008, 0.007
0.007 0.04
0.002 0.002
0.03 0.02
contained in birth control pills. A recent web search of currently
available birth control products (6/2019) finds the lowest dose of
EE2 in (Lo)estrin FE R
, containing 10 µg of EE2/day, or 1/3 of
the “safe dose” for E2. While this product has not been tested in
our laboratory, pure EE2 has been evaluated and determined to
be equivalent to 17β-E2 by E-Screen (data not shown). In this
context, one tofu burger has 1/10th of the E2Eqs theoretically
present in this product.
It is not clear what level of consumption of E2 or phytoestro-
gens confers risk or benefit (Zaheer et al., 2017 review). Some
scientists have attempted to link the consumption of milk with
early onset of puberty (Maruyama, Oshima, & Ohyama, 2010).
The bioavailability of dietary phytoestrogens, such as genistein,
has been established in humans (Xu, Wang, Murphy, & Hendrich,
2000). Though soy supplements containing genistein were found
to alter androgen receptor expression, they did not alter measured
circulating hormone concentrations of men (Hamilton-Reeves,
Rebello, Thomas, Slaton, & Kurzer, 2007). In contrast, hormone
concentrations of women receiving genistein supplements were
altered (Petrakis et al., 1996).
Table 1 provides a theoretical estimate of circulating E2Eqs post-
consumption of one serving of the foods assayed. E2Eq/serving
was multiplied by 10% as a nominal value for absorption and
divided by the total blood volume (women, men, and children,
Blood Volume Calculator) to obtain E2Eq/mL. These estimates
are listed next to circulating estradiol concentrations. Obvious
caveats include differential rates of absorption for estrogens, phy-
toestrogens, and zearalenone; assuming an instantaneous absorp-
tion of total “dose” and no acknowledgment of metabolism or
elimination of estrogenic compounds. Absorption of genistein was
clearly documented in a study of women consuming diets contain-
ing genistein, in which approximately 18% of dietary intake was
recovered in the urine (Xu et al., 2000). Using this “one size fits
all” approach, tofu was the only food examined that might be ex-
pected to contribute significantly to the estrogenicity of circulating
serum. In concordance with modeled results, circulating plasma
concentrations of genistein in 4-month old infants on soy-based
infant formula of 684 versus 3 ng/mL in plasma of infants fed cow’s
milk (Setchell, Zimmer-Nechemias, Cai, & Heubi, 1997). By
Estrogenic equivalents per food serving . . .
E-Screen, 684 ng/mL of genistein would result in 103 pg/mL
E2 Eq, compared to circulating E2 concentrations reported in in-
fants of ࣘ25 pg/mL (Konforte et al., 2013).
Conclusion
The E-Screen can be a useful in vitro tool as a first step in evalu-
ating potentially estrogenic foods prior to in vivo studies. We found
that estimates of estrogenic activity of extracts based on chemical
concentrations of the most estrogenic compounds in a food multi-
plied by its E2 equivalency factor (assessed by E-Screen on the pure
chemical) were similar to what was found by E-Screen of food ex-
tracts. These findings are in agreement with literature reports of E2
concentrations, and increase the credibility of the data set, though
limited in both its size and representative nature. Application of
actual in vivo human bioavailability data for estrogenic compounds
in foodstuffs and knowledge of their human metabolism would
improve the model’s crude theoretical estimates of blood E2Eqs
concentrations for comparison to endogenous estradiol circulating
blood concentrations for children and adults. This study clearly in-
dicates that, by E-Screen, the estrogenic activity of milk, rice, and
ground beef is much lower than tofu.
Acknowledgments
Cass Clay Creamery provided vitamin supplements used in the
fortification of commercial milk for E-Screen testing. Composi-
tional analysis was kindly provided by Drs. Michael Bukowski and
Igathinathane Cannayen. Mr. Patrick Harland was integral to the
analysis of the ground beef samples.
Author Contributions
Nancy W. Shappell designed, conducted, and analyzed exper-
iments, and wrote the manuscript. Eric P. Berg and James D.
Magolski provided initial composite beef samples, their analyses,
stimulus for work, and manuscript review. Lloyd O. Billey was es-
sential to the conduct of experiments and manuscript preparation.
References
Al-Ansari, A. M., Saleem, A., Kimpe, L. E., Trudeau, V. L., & Blias, J. M. (2011). The develop-
ment of an optimized sample preparation for trace level detection of 17α-ethinylestradiol and
estrone in whole fish tissue. Journal of Chromatography B: Biomedical Sciences and Applications,
879, 3649–3652.
Blair, A. (2015). Hormones in beef: Myth vs, fact. Retrieved from https://www.drovers.
com/article/hormones-in-beef-myth-vs-fact-naa-university-news-release/
Blood Volume Calculator. (2019). Retrieved from http://www.easycalculation.com/medical/
blood-volume.php
Braekevelt, E., Lau, B. P., Tague, B., Popovic, S., & Tittlemier, S. A. (2011). Effect of cooking on
concentrations of ß-estradiol and metabolites in model matrices and beef. Journal of Agricultural
and Food Chemistry, 59, 915–920.
Brett, K. M. & Chang, Y. (2001). Hormone replacement therapy: knowledge and use in the United
States. Hyattsville, MD: National Center for Health Statistics.
Bryant, T. C., Engle, T. E., Galyean, M. L., Wagner, J. D., Tatum, J. D., Anthony, R. V.,
& Laudert, S. B. (2010). Effects of ractopamine and trenbolone acetate implants with or
without estradiol on metabolites in feedlot steers and heifers growth performance, carcass
characteristics, adipogenic enzyme activity, and blood. Journal of Animal Science, 88, 4102–
4119.
Capriotti, A. L., Cavaliere, C., Foglia, P., Samperi, R., Stampachiacchiere, S., Ventura, S., &
Lagana, A. (2015). Ultra-high performance liquid chromatography-tandem mass spectrometry
for the analysis of free and conjugated natural estrogens in cow milk without deconjugation.
Analytic and Bioanalytic Chemistry, 407, 1705–1719.
CDC. (2012). DEC update: Consumers. Retrieved from https://www.cdc.gov/des/consumers/
download/index.html
Chen, C., Mi, X., Yuan, Y., Chen, G., Ren, L., Wang, K., . . . Qian, Y. (2014). A preliminary
risk assessment of potential exposure to naturally occurring estrogens from Beijing (China)
market milk products. Food and Chemical Toxicology, 71, 74–80.
Code of Federal Regulations. (2011). CFR 556.240, Estradiol and related esters. Title 21. Vol.
11. Retrieved from https://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfcfr/CFRSearch.
cfm?fr=556.240
Courant, F., Antignac, J.-P., Maume, D., Monteau, F., Andre, F., & Le Bizec, B. (2007). Deter-
mination of naturally occurring oestrogens and androgens in retail samples of milk and eggs.
Food Additives & Contaminants, 24, 1358–1366.
Furnari, C., Maroun, D., Gyawali, S., Snyder, B. W., & Davis, A. M. (2012). Lack of biologically
active estrogens in commercial milk. Journal of Dairy Science, 95, 9–14.
Goyon, A., Cai, J. Z., Kraehenbuehl, K., Hartmann, C., Shao, B., & Mottier, P. (2016). De-
termination of steroid hormones in bovine milk by LC-MS/MS and their levels in Swiss
Holstein cow milk. Food Additives and Contaminants, 33, 804–816.
Hamilton-Reeves, J. M., Rebello, S. A., Thomas, W., Slaton, J. W., & Kurzer, M. S. (2007).
Isoflavone-rich soy protein isolate suppresses androgen receptor expression without altering
estrogen receptor-ß expression or serum hormonal profiles in men at high risk of prostate
cancer. Journal of Nutrition, 137, 1769–1775.
Hartmannn, S., Lacorn, M., & Steinhart, H. (1998). Natural occurrence of steroid hormones in
food. Food Chemistry, 62, 7–20.
Impens, S., De Wasch, K., Cornelis, M., & De Brabander, H. F. (2002). Analysis on residues of
estrogens, gestagens and androgens in kidney fat and meat with gas chromatography-tandem
mass spectrometry. Journal of Chromatography A, 970, 235–247.
Joint FAO/WHO Expert Committee on Food Additives. (2000a). Evaluation of certain veterinary
drug residues in food. 53rd Report of the Joint FAO/WHO Expert Committee on Food
Additives, Series 44. Zearalenone. Geneva, Switzerland: World Health Organization. Table 5,
Section 6.
Joint FAO/WHO Expert Committee on Food Additives. (2000b). Evaluation of certain veterinary
drug residues in food. 52nd Report of the Joint FAO/WHO Expert Committee on Food
Additives, Series 43. Estradiol-17 ß, progesterone, and testosterone. Geneva, Switzerland:
World Health Organization. 49.
Konforte, D., Shea, J. L., Kyriakopoulou, L., Colantonio, D., Cohen, A. H., Shaw, J., . . . Adeli,
K. (2013). Complex biological pattern of fertility hormones in children and adolescents: A
study of healthy children from the CALIPER cohort and establishment of pediatric reference
intervals. Clinical Chemistry, 59(8), 1215–1227.
Kushinsky, S. (1983). Safety aspects of the use of cattle implants contaning natural steroids.
International Symposium on Safety Evaluation of Animal Drug Residues, Berlin.
Magolski, J. D., Shappell, N. W., Vonnahme, K. A., Anderson, G. M., Newman, D. J., & Berg,
E. P. (2014). Consumption of ground beef obtained from cattle that had received steroidal
growth-promotants did not trigger early onset of estrus in pre-pubertal pigs. Journal of Nutrition,
144, 1718–1724.
Maruyama, K., Oshima, T., & Ohyama, K. (2010). Exposure to exogenous estrogen through
intake of commercial milk produced from pregnant cows. Pediatrics International, 52(1), 33–38.
Mayo ClinicBlood estradiol values. Retrieved from https://www.mayomedicallaboratories.com/
test-catalog/Clinical±and±Interpretive/84230
Murphy, P. A., Song, T., Buseman, G., Barua, K., Beecher, G. R., Trainer, D., & Holden, J.
(1999). Isoflavones in retail and institutional soy foods. Journal of Agricultural and Food Chemistry,
47(7), 2697–2704.
National Cancer Institute. (2011). Diethylstilbestrol (DES) and cancer. Bethesda, MD: U. S. National
Institutes of Health. Retrieved from http://www.cancer.gov/cancertopics/factsheet/risk/des
Pape-Zambito, D. A., Magliaro, A. L., & Kensinger, R. S. (2007). Concentrations of 17ß-
estradiol in Holstein whole milk. Journal of Dairy Science, 90(7), 3308–3313.
Pape-Zambito, D. A., Magliaro, A. L., & Kensinger, R. S. (2008). 17ß-Estradiol and estrone
concentrations in plasma and milk during bovine pregnancy. Journal of Dairy Science, 91(1),
127–135.
Petrakis, N. L., Barnes, S., King, E. B., Lowenstein, J., Wiencke, J., Lee, M. M., . . . Cow-
ard, L. (1996). Stimulatory influence of soy protein isolate on breast secretion in pre- and
postmenopausal women. Cancer Epidemiology, Biomarkers & Prevention, 5, 785–794.
Rassmussen, T. H., & Nielsen, J. B. (2002). Critical parameters in the MCF-7 cell proliferation
bioassay (E-Screen). Biomarkers, 7(4), 322–336.
Raun, A. P., & Preston, R. L. (2002). History of diethylstilbesterol use in cattle. ASAS Organization.
Retrieved from http://www.asas.org/docs/publications/raunhist.pdf?sfvrsn=0
Seo, J., Kim, H. Y., Chung, B. C., & Hong, J. (2005). Simultaneous determination of anabolic
steroids and synthetic hormones in meat by freezing-lipid filtration, solid-phase extraction
and gas chromatography-mass spectrometry. Journal of Chromatography A, 1067(1–2), 303–
309.
Setchell, K. D., Zimmer-Nechemias, L., Cai, J., & Heubi, J. E. (1997). Exposure of infants to
phyto-oestrogens from soy-based infant formula. Lancet, 350, 23–27.
Shappell, N. W. (2006). Estrogenic activity in the environment: Municipal wastewater effluent,
river, ponds, and wetlands. Journal of Environmental Quality, 35, 122–132.
Shappell, N. W., Mostrom, M. S., & Lenneman, E. M. (2012). E-Screen evaluation of sugar beet
feedstuffs in a case of reduced embryo transfer efficiencies in cattle: The role of phytoestrogens
and zearalenone. In Vitro Cellular & Developmental Biology - Animal, 48(4), 216–228.
Soto, A. M., Sonnenschein, C., Chung, K. L., Fernandez, M. F., Olea, N., & Serrano, F.
O. (1995). The E-SCREEN assay as a tool to identify estrogens: An update on estrogenic
environmental pollutants. Environmental Health Perspectives, 103(Suppl. 7), 113–122.
Stephany, R. W. (2010). Hormonal growth promoting agents in food producing animals. Hand-
book of Experimental Pharmacology, 195, 355–367.
Stob, M., Andrews, F. N., Zarrow, M. X., & Beeson, W. M. (1954). Estrogenic activity of the
meat of cattle, sheep and poultry following treatment with synthetic estrogens and proges-
terone. Journal of Animal Science, 13(1), 138–151.
Stypula-Trebas, S., Minta, M., Radko, L., & Zmudzki, J. (2015). Application of the yeast-based
reporter gene bioassay for the assessment of estrogenic activity in cow’s milk from Poland.
Environmental Toxicology and Pharmacology, 40, 876–885.
Synovex R
Plus. (1996). Synovex R
Plus – Original approval (trenbolone acetate and estradiol benzoate).
NADA 141-043. Silver Spring, MD: U.S. Food and Drug Administration.
Tate, P. L., Bibb, R., & Larcom, L. L. (2011). Milk stimulates growth of prostate cancer cells in
culture. Nutrition and Cancer, 63, 1361–1366.
Vicini, J., Etherton, T., Kris-Etherton, P., Ballam, J., Denham, S., Staub, R., . . . Lucy,
M. (2008). Survey of retail milk composition as affected by label claims regarding farm-
management practices. Journal of the American Dietetic Association, 108(7), 1198–1203.
Wolford, S. T., & Argoudelis, C. J. (1979). Measurement of estrogens in cow’s milk, human milk,
and dairy products. Journal of Dairy Science, 62(9), 1458–1463.
Xu, X., Wang, H. J., Murphy, P. A., & Hendrich, S. (2000). Neither background diet nor type
of soy food affects short-term isoflavone bioavailability in women. Journal of Nutrition, 130,
798–801.
Zaheer, K., & Akhtar, M. H. (2017). An updated review of dietary isoflavones: Nutrition,
processing, bioavailability and impacts on human health. Critical Reviews in Food Science and
Nutrition, 57, 1280–1293.
Toxicology & Chemical
Food Safety
Vol. 84, Iss. 12, 2019 r Journal of Food Science 3883
 Estrogenic equivalents per food serving . . .

Supporting Information Table S5. Concentrations of genistin and genistein by LC-

a b
Additional supporting information may be found online in the MS/MS and predicted and observed E2 Eqs of tofu .
Supporting Information section at the end of the article. Figure S1. Representative proliferative response of MCF-7 cells
Table S1. Compositional analysis (%) of cooked ground beef and to standard curves of various pure compounds. Cells were incu-
tofu burgers. bated with pure compounds for 5 days, fixed, protein stained,
solubilized, and absorbance read at 490 nm, n = 5 wells. Estra-
Table S2. Conditions for LC-MS/MS analysis of compounds. diol equivalents were calculated based on standard curve re-
Table S3. Information on cattle that ground beef samples were sponses from a minimum of three separate experiments, with the
prepared from. Some information was not available (na). exception of trenbolone (two experiments). Figure was previ-
ously published in Shappell et al. (2012) without the trenbolone
Table S4. Estrogenicity of various foods and PremarinTM assayed data.
by E-Screen.
Toxicology & Chemical
Food Safety

3884 Journal of Food Science r Vol. 84, Iss. 12, 2019

View publication stats