Isoflavones in human breast milk and other biological fluids

Isoflavones in human breast milk and other biological fluids1–3
Adrian A Franke, Laurie J Custer, and Yuichiro Tanaka
ABSTRACT
We established a method for using HPLC and
ciated with the beneficial effects of soy consumption, there is
diode-array ultraviolet scanning to quantitate soy isoflavonoids
great interest in rapid, accurate, and affordable methods to quan-
in foods and in human plasma, urine, and breast milk. The ana-
titate these analytes in foods and biological material. In the past,
lytes occurring as glycoside conjugates were hydrolyzed enzy-
gas chromatography–mass spectrometry (GC-MS) was favored
matically before HPLC analysis if extracted from biological
for the analysis of soy isoflavones and their metabolites in
matrices or were subjected to direct HPLC analysis after extrac-
human plasma (12), urine (13–16), and feces (17, 18). In recent
tion from foods. We monitored the isoflavones daidzein, genis-
years, however, the use of HPLC for quantitating isoflavonoids
tein, glycitein, formononetin, and biochanin-A and their mam-
in these body fluids was introduced (19–22). This method allows
malian metabolites equol and O-desmethylangolensin in human
a variety of isoflavonoid phytoestrogens, including aglycones
Downloaded from
plasma, urine, and breast milk. Analytes were identified by
and conjugated analytes, to be measured in one operation. Com-
absorbance patterns, fluorometric and electrochemical detection,
pared with GC-MS, HPLC requires fewer steps for sample
and comparison with internal and external standards. In addition,
preparation and analysis, less technician time, and less expensive
we identified analytes by using gas chromatography–mass spec-
instrumentation. In studies of dietary intervention with soy,
trometry after trimethylsilylation. The HPLC method was also
HPLC (19) proved to be as accurate as GC-MS (13, 15) in meas-
used to measure concentrations of isoflavones and their gluco-
uring urinary isoflavone concentrations.
www.ajcn.org
side conjugates in various soy-based infant formulas. Total
To improve methods for assessment of the role of dietary soy
isoflavone concentrations varied between 155 and 281 mg/kg.
in promoting human health, we established a technique that com-
After one woman received a moderate challenge with 20 g roast-
bines HPLC and diode-array ultraviolet scanning to quantitate
ed soybeans (equivalent to 37 mg isoflavones), we detected mean
soy isoflavonoids in foods and in human plasma (21), urine (19),
by on August 14, 2009
total isoflavone concentrations of ?2.0 ?mol/L in plasma, 0.2
and breast milk (23). The purpose of the present study was to
?mol/L in breast milk, and 3.0 ?mol/h in urine. According to our
quantitate glucuronide and sulfate conjugates of daidzein and
measurements, with adjustment for body weight, isoflavonoid
genistein in human plasma, urine, and breast milk by using this
exposure is 4–6 times higher in infants fed soy-based formula
HPLC method. The HPLC method was also used to monitor
than in adults eating a diet rich in soyfoods (?30 g/d). Implica-
isoflavones and their glucoside conjugates in soy-based infant
tions of the presented results for the potential cancer-preventing
formulas. We discuss the implications of our results in relation to
activity of isoflavones by exposing newborn infants to these phy-
the potential for cancer prevention by exposure of newborn
tochemicals are discussed.
Am J Clin Nutr
infants to isoflavones.
1998;68(suppl):1466S–73S.
KEY WORDS
Daidzein, genistein, glycitein, isoflavonoids,
SUBJECTS AND METHODS
soy, breast milk, urine, plasma, soy-based infant formula, HPLC
Apparatus
We performed HPLC analyses on a system Gold chromatograph
INTRODUCTION
with an autosampler (model 507), a dual-channel diode-array
Soy consumption is suggested to contribute to the prevention
detector (model 168) (both from Beckman, Fullerton, CA), a ?uo-
of chronic diseases, including cardiovascular disorders, osteo-
rescence detector (model FD100; GTI/SpectroVision, Concord,
porosis, and cancer. New findings in this regard are presented in
MA), and an electrochemical 5011 coulometric cell detector
detail in these conference proceedings. Recent animal experi-
(Coulochem II-5200; ESA, Bedford, MA). Absorbance readings
ments show that only 3 doses of genistein given to newborn, pre-
were obtained with a spectrophotometer (DU-62, Beckman). Evap-
pubertal, or perinatal rats can reduce the incidence and number
of breast tumors (1). In addition, genistein in combination with
1 From the Cancer Research Center of Hawaii, Honolulu.
daidzein produces a highly significant reduction in neoplastic
2 Supported by American Cancer Society Institutional Grant IRG 190 and
transformation in a rodent cell model (2).
a grant from the Hawaii Chamber of Commerce.
Soyfoods containing ?0.2–0.3% daidzein and genistein
3 Address reprint requests to AA Franke, Cancer Research Center of
(3–10) were identified as the predominant source of human
Hawaii, 1236 Lauhala Street, Honolulu, HI 96813. E-mail: adrian@
exposure to isoflavones (3, 11). Because isoflavonoids are asso-
crch.hawaii.edu.
1466S
Am J Clin Nutr 1998;68(suppl):1466S–73S. Printed in USA. © 1998 American Society for Clinical Nutrition

---

ISOFLAVONES IN HUMAN BIOLOGICAL FLUIDS
1467S
oration was performed with a speed vacuum (AS 160; Savant,
in 25 mL of 80% aqueous methanol containing 20 ppm flavone
Farmingdale, NY) at room temperature. We performed GC analy-
as an internal standard. Next, the mixture was stirred for 2 h in a
sis with a mass spectrometer (model 5890; Hewlett-Packard, Wilm-
sealed container at room temperature. After centrifugation at 850
ington, DE), using a mass selective detector (model 5971A,
? g for 5 min at ambient temperature, a clear aliquot of the solu-
Hewlett-Packard) and electron impact ionization at 70 eV. Solid-
tion was diluted 1:1 with acetate buffer (0.2 mol/L, pH 4); 20 ?L
phase, C
reversed-phase extraction columns were obtained from
of this solution, which contained the isoflavone conjugates and
18
PGC Scienti?c (Gaithersburg, MD).
aglycones, was injected into the HPLC system.
Chemicals
Enzymatic hydrolysis and extraction of isoflavones from
urine
Methanol, hydrochloric acid, acetic acid, 96% ethanol,
dimethyl sulfoxide, ethyl acetate, and all solvents used for HPLC
For extraction of isoflavones from urine, we processed the
and absorbance readings were analytic grade or HPLC grade
urine as described previously (19, 21). In brief, 20 mL clear
from Fisher Scientific (Fair Lawn, NJ). Butylated hydroxy-
urine was mixed with 5.0 mL acetate buffer (0.2 mol/L, pH 4)
toluene, sodium acetate, and genistin were purchased from
and 200 ?L flavone (60 ppm in 96% ethanol) as the internal stan-
Sigma Chemical Co (St Louis). Daidzein and genistein were
dard and passed through a preconditioned C
solid-phase extrac-
18
obtained from ICN (Costa Mesa, CA), flavone was obtained
tion column. The mixture was then washed with 2 mL acetate
from Aldrich (Milwaukee), and coumestrol was obtained from
buffer and the analytes were eluted with 100% methanol. The
Serva (Westbury,
NY). ?-Glucuronidase isolated from
eluate was dried under reduced pressure at room temperature,
Escherichia coli (200 ? 103 U/L) and arylsulfatase isolated from
redissolved in 1.0 mL phosphate buffer (0.2 mol/L, pH 7.0),
Helix pomatia (1–5 U/mL) were purchased from Boehringer
mixed thoroughly with 10 ?L ?-glucuronidase (24) and 10 ?L
Mannheim (Indianapolis). Equol and O-desmethylangolensin
arylsulfatase, and incubated for 1 h at 37 ?C. Subsequently, we
were purchased from the University of Helsinki.
inactivated the enzymes of the hydrolyzed samples by adding
Downloaded from
0.98 mL of 100% methanol. Samples were analyzed immedi-
Subject
ately or stored at ?20 ?C and analyzed by HPLC after equilibra-
One woman of average height and weight who was in week 15
tion to room temperature, vortex mixing, and centrifugation for
of the postpartum period participated in the study. She did not
5 min at 850 ? g and 8 ?C. Additional concentration was
smoke; took no medication, hormones, or dietary supplements;
achieved by partitioning the isoflavones from hydrolyzed sam-
had no specific dietary patterns (eg, vegetarian); and was in good
ples into ethyl acetate, combining the organic phases, and drying
www.ajcn.org
health. This woman donated breast milk, blood, and urine for the
under nitrogen at room temperature. We redissolved the residue
present study after signing an informed consent form. During the
in 150 ?L mobile phase solvent and 50 ?L acetate buffer (0.2
study, the subject did not consume any alcohol and maintained
mol/L, pH 4) before injecting 20 ?L of the solution into the
her usual diet except for the intake of 5, 10, and 20 g roasted soy-
HPLC system.
by on August 14, 2009
beans containing 0.85 mg daidzein/kg and 1.1 mg genistein/kg
(3) during the dietary intervention period. The soy doses were
Enzymatic hydrolysis and extraction of isoflavones from
added to the women’s usual self-selected diet. All procedures of
breast milk
the protocol were in accordance with the ethical standards of the
As reported previously (23), we mixed 2–4 mL breast milk
Helsinki Declaration of 1975 as revised in 1983.
equilibrated to room temperature with 25 ?L flavone (120 ppm
in 96% ethanol), 50 ?L ?-glucuronidase, and 50 ?L arylsulfa-
Collection and handling of human milk, blood, and urine
tase. The mixture was stirred for 1 h at 37 ?C. This sample was
Each time after the woman nursed her infant, milk was col-
extracted 3 times with 2 mL ethyl acetate (certified by the Amer-
lected from the breast that was not used for breast-feeding. Col-
ican Chemical Society) and the organic phases were combined
lection included entire expression of the breast. We stored milk
after centrifugation at 850 ? g for 5 min at ambient temperature,
samples in plastic vials at 6 to ?4 ?C. Isoflavones in milk were
followed by drying under nitrogen at room temperature. The dry
stable for ? 5 d when the milk was kept within this temperature
extract was redissolved in 150 ?L methanol by vortex mixing.
range (23).
Then we added 50 ?L acetate buffer (0.2 mol/L, pH 4). After
After the subject fasted overnight, blood samples were drawn
centrifugation for 20 min at 850 ? g at 8 ?C, 20 ?L of the clear
into evacuated containers containing heparin and kept on ice until
sample was injected into the HPLC system.
centrifuged at 850 ? g for 30 min at 4 ?C. The supernatant plasma
was either analyzed immediately or stored at ?70 ?C (21).
Enzymatic hydrolysis and extraction of isoflavones from
Urine samples were stored in disposable bottles containing
plasma
0.2 g boric acid and 0.1 g ascorbic acid to control bacterial con-
We mixed 1.0 mL plasma equilibrated to room temperature
tamination and degradation of analytes (19). After we mixed
with 0.25 mL triethylamine acetate (0.5 mol/L, pH 7.0), 80 ?L
each urine sample and measured the volume, we transferred the
?-glucuronidase, 80 ?L arylsulfatase, and 20 ?L flavone (120
sample into a 50-mL disposable plastic tube and stored it at ?4
ppm in 96% ethanol). The mixture was stirred for 17 h at 37 ?C
to ?20 ?C. Each urine volume, time of urine collection, and time
in a sealed container as reported previously (21). After adding
of previous voiding was recorded so that values could be
0.25 mL of 10% aqueous trichloroacetic acid, we extracted the
adjusted for urine concentration.
mixture 3 times with 2 mL ethyl acetate and dried the combined
organic phases under nitrogen. The residue was redissolved in
Extraction of isoflavones from soy-based infant formula
150 ?L methanol by vortex mixing. Then we added 50 ?L
Soy-based infant formulas were purchased from local stores
acetate buffer (0.2 mol/L, pH 4) and sonicated the mixture for 30
and extracted by sonicating 0.5 g dry formula powder for 10 min
s. After centrifugation of 20 ?L at 4 ?C for 5 min, the clear sam-

---

1468S
FRANKE ET AL
Downloaded from
FIGURE 1. High-performance liquid chromatograms and diode-array ultraviolet scans of daidzein and genistein obtained from human breast milk.
Milk was collected 18 h after a challenge with 20 g roasted soybeans and extracted as described in the Methods. Flavone was used as the internal stan-
dard and eluted after 31 min (not shown). Amounts shown correspond to 3.7 pmol daidzein and 5.4 pmol genistein. We did not detect glycitein, equol,
or O-desmethylangolensin. Absorbance scans obtained from standards for daidzein and genistein (solid lines) are identical to those obtained from
HPLC peaks (dotted lines), indicating the presence of daidzein and genistein in this milk extract.
www.ajcn.org
ple was injected into the HPLC system.
adding methanol to give stock solutions of 2–5 mmol/L. We dis-
carded compounds with < 95% purity as determined by HPLC
Trimethylsilylation
analysis. The concentration of the stock solutions was deter-
Dry milk extracts or crystalline standards were dissolved in
mined by absorbance readings as reported previously (19). Con-
by on August 14, 2009
0.1 mL trimethylsilyltrifluoracetamide-imidazole (100:2, vol:wt)
centrations of analytes in milk and urine were calculated from
and incubated for 15 min at 60 ?C before GC-MS analysis (24).
peak areas that we determined after HPLC analyses by using the
slopes of the calibration curves obtained from serial dilutions of
Chromatographic conditions
standard solutions. Milk and plasma concentrations were
All HPLC analyses were performed on a NovaPak C
expressed as nmol/L and values were adjusted for recovery of the
18
reversed-phase column [150 ? 3.9 mm; inside diameter: 4 ?m
internal standard. Urinary excretion rates were expressed as
(Waters, Milford, MA)] coupled to an Adsorbosphere C
direct-
nmol/h after adjustment for the time period between the urine
18
connect guard column [10 ? 4.6 mm; inside diameter: 5 ?m
collection and previous voiding (in hours), urine volume (in mil-
(Alltech, Deerfield, IL)]. We performed elution at a flow rate of
liliters), and recovery of the internal standard.
0.8 mL/min with the following linear gradient: A, acetic acid to
water (10:90, by vol); B, methanol-acetonitrile-dichloromethane
(10:5:1, by vol); and B in A (by vol) at 5% for 5 min, from 5%
RESULTS
to 45% in 20 min, from 45% to 70% in 6 min, and from 70% to
Analytes were routinely identified by retention times in vari-
5% in 3 min. Signals were scanned between 190 and 400 nm.
ous HPLC systems and by diode-array absorption patterns (Fig-
We performed gas chromatography by injecting 3 ?L of the
ure 1 and Figure 2). The HPLC system separated completely
trimethylsilylated sample onto a Hewlett-Packard P Ultra-1 cap-
parent and metabolized isoflavonoids such as daidzein (16.8
illary column (inside diameter: 17 m ? 0.2 mm; film thickness:
min), glycitein (17.9 min), genistein (20.1 min) and para-ethyl-
0.11 ?m); helium was used as the carrier gas at a flow rate of 1.0
phenol (22.1 min), equol (20.8 min) and O-desmethylan-
mL/min with a 1:10 flow split. The following temperature pro-
golensin (23.2 min), and mammalian lignans including entero-
gram was used: initial temperature, 180 ?C; rate, 10 ?C/min; and
diol (18.9 min) and enterolactone (22.6 min). Fluorometric and
final temperature, 320 ?C. Signals were registered in the selected
coulometric detection (19, 21, 25–27) confirmed the presence of
ion monitoring (SIM) mode and the following masses were
daidzein and genistein in milk, with detection limits being 2–6-
determined after analysis of standards: daidzein: mass-to-charge
fold lower than those with monitoring by ultraviolet absorbance.
ratio (m/z) 398, 383; genistein: m/z 471, 399, 228; equol: m/z
Finally, GC-MS-SIM analysis of trimethylsilylated milk
386, 192; and O-desmethylangolensin: m/z 459, 281.
extracts revealed GC retention times and mass fragmentation
patterns identical to those for daidzein and genistein standards
Standard solutions and calibration curves
(18, 23), confirming the presence of these isoflavones in human
Standard stock solutions were prepared by dissolving 1–3 mg
milk.
of the crystalline compound in 20 ?L dimethyl sulfoxide and
The soybean intake challenge led to a rapid and dose-depen-

---

ISOFLAVONES IN HUMAN BIOLOGICAL FLUIDS
1469S
Downloaded from
www.ajcn.org
by on August 14, 2009
FIGURE 2. High-performance liquid chromatograms of isoflavonoids obtained from extracts of human urine and plasma and compared with that
of standards for isoflavonoids. DE, daidzein; GLYE, glycitein; GE, genistein; DMA, O-desmethylangolensin; EQ, equol; COM, coumestrol.
dent response in isoflavone concentrations in breast milk (Fig-
mula. Patterns of conjugation and isoflavone type followed those
ure 3). Our results are in excellent agreement with results
expected for soy protein isolate (6, 8, 28).
reported recently by GC-MS analysis (28). Maximum milk con-
centrations were reached 10–14 h after soy intake and baseline
concentrations were estimated to be reached 2–4 d later, depend-
DISCUSSION
ing on the dose.
We examined the presence of isoflavonoids in breast milk to
Extending our previous work on the measurement of
assess exposure of breast-fed infants to isoflavones, phytochem-
isoflavone concentrations in food items (3, 21), we analyzed
icals that may contribute to cancer prevention (23). Recovery of
concentrations of isoflavone aglycones and individual isoflavone
analytes was low for both solid-phase extraction methods and
conjugates from various soy-protein-based infant formulas.
phase-separation techniques with various organic solvents. This
Again, HPLC peaks of extracts (Figure 4) were identified by
was probably due to the high protein content and the emulsion-
comparison of diode-array absorbance patterns (5, 21). Ultravio-
like nature of milk. However, phase separation with ethyl acetate
let scans of conjugated isoflavones showed patterns identical to
resulted in selective concentration of the analytes and led to
those of their respective aglycones (Figure 4). This result is in
recovery that was superior to that achieved with any of the other
excellent agreement with the finding that glycosylation at the C-
organic solvents used. Isolation with ethyl acetate resulted in
7 hydroxyl position of the flavonoid molecule does not affect the
high precision (CV: 5–11%) and spiking recoveries (88–99%) of
flavonoid chromophore (29). Total isoflavone amounts (Table 1)
isoflavone aglycones from breast milk and did not coextract gly-
varied between 155 and 281 ?g/g depending on the brand of for-
cosidic conjugates (23). To include conjugates in the assay, we

---

1470S
FRANKE ET AL
Downloaded from
FIGURE 3. Daidzein and genistein concentrations in human breast milk and urine after a challenge with 5, 10, and 20 g roasted soybeans given at
0, 24, and 72 h. Concentrations were determined by HPLC with diode-array ultraviolet scanning. Milk samples were collected at the time of feeding.
Six overnight urine samples and 2 additional samples, collected 12 h before and 87 h after the first soybean intake, were obtained to monitor urinary
isoflavone excretion.
www.ajcn.org
had to perform hydrolysis before extraction (24, 25). The lack of
to enterohepatic circulation (37), a process that commonly
detectable isoflavones after extraction when hydrolysis was not
occurs after exposure to flavonoid agents (38, 39). Enterohepatic
performed suggests that all isoflavones occur as glucuronide or
circulation may also explain the biphasic pattern observed in
by on August 14, 2009
sulfate conjugates in human milk.
human milk. Using mass spectrometric techniques, we detected
The appearance of isoflavones in milk followed their appear-
glycitein and the isoflavone metabolites equol and O-
ance in urine, with a slight delay (Figure 3). This result is in
desmethylangolensin in urine but not in breast milk (23). This
good agreement with findings for other micronutrients or drugs
finding suggests either that the quantities of these agents in
for which urinary excretion precedes secretion in breast milk
breast milk are below detection limits or that the agents were not
(30). Preferential excretion of the metabolites over the parent
secreted into human milk when the woman who participated in
isoflavones is suggested by the ratio of the isoflavone to its
the study consumed soybeans. The inability to detect glycitein in
metabolite, which was previously observed to be much higher in
breast milk is most likely due to low dietary glycitein exposure
plasma than in urine or feces (12, 18). Because milk is likely to
resulting from the low concentration in soybeans (5–10% of total
be produced from blood by passive secretory processes (30), pat-
isoflavones) (4, 6, 9, 21).
terns in milk probably reflect those in plasma.
Isoflavone concentrations in 4 different soy-based infant for-
Concentrations of genistein conjugates in milk were higher
mulas varied by a factor of 2 (Table 1), in part because of the dif-
than those of daidzein conjugates (Figure 3). We made a similar
ferent amount of soy protein isolate in these formulas. More
observation for concentrations in plasma: 15 h after the intake of
important, however, considering the influence of environmental
20 g soybeans, the concentration of genistein was 1039 nmol/L
conditions and genetic dispositions on the isoflavone content of
and that of daidzein was 481 nmol/L (data not shown). In con-
soybeans (7, 9, 40), this variability was probably due to the dif-
trast, urinary concentrations of genistein conjugates were lower
ferent sources of soybeans used in these formulas. As determined
than those of daidzein conjugates, which is consistent with find-
by our analyses, conjugation patterns of soy isoflavones in infant
ings in other studies (14–16, 19, 20, 31–33). The explanation
formulas were not significantly different from those in tradi-
may be that the higher polarity of daidzein conjugates favors uri-
tional soyfoods (21), with malonates (32–43%) and glucosides
nary excretion, leading to increasing ratios of genistein to
(37–52%) dominating over acetates (6–7%) and aglycones
daidzein in the blood.
(9–13%). As reported previously (10), soy-based infant formulas
A distinct biphasic pattern of milk concentrations was
contain appreciable amounts of isoflavones. Our analyses of 4
observed, especially after the challenge with 20 g soybeans (Fig-
commonly available soy-based infant formulas revealed average
ure 3). This phenomenon was also reported when urinary (19)
total concentrations of daidzein of 77 ± 25 ?g/g, of glycitein of
and plasma (34) isoflavone patterns were monitored in human
18 ± 5 ?g/g, and of genistein of 122 ± 35 ?g/g. Considering that
subjects after soy dietary intervention or in animals after expo-
an average infant consumes 1 L liquid food and weighs 4.5 kg,
sure to various flavonoids (35, 36). The findings of a study in
by preparing the soy-based formula according to the directions
rats suggest that the biphasic pattern of isoflavones may be due
on the label (8.7 g powder per 60 mL water), which we discov-

---

ISOFLAVONES IN HUMAN BIOLOGICAL FLUIDS
1471S
Downloaded from
www.ajcn.org
FIGURE 4. High-performance liquid chromatogram and diode-array ultraviolet scans of daidzein, glycitein, and genistein and their malonylgluco-
side, acetylglucoside, and glucoside conjugates obtained from a soy-based infant formula extract. D, daidzein-7-O-glucoside (daidzin); GLY, glycitein-
7-O-glucoside (glycitin); G, genistein-7-O-glucoside (genistin); D-MAL, daidzein-7-O-malonyl glucoside; GLY-Mal, glycitein-7-O-malonyl
by on August 14, 2009
glucoside; G-Mal, genistein-7-O-malonyl glucoside; D-Ac, daidzein-7-O-acetyl glucoside; G-Ac, genistein-7-O-acetyl glucoside; DE, daidzein; GE,
genistein; GLYE, glycitein.
ered to contain on average 0.21 mg isoflavones/g, the daily aver-
isoflavones by using HPLC and diode-array ultraviolet scanning
age intake of isoflavones is ?7 mg/kg. After adjustment for body
is suitable for use in future epidemiologic and clinical trials to
weight, this intake is 4–6 times greater than that of an adult con-
assess the role of soy or isoflavonoids in preventing chronic dis-
suming soyfoods regularly (30 g soy protein/d). Whether this rel-
ease. After a healthy woman was given a moderate challenge
atively high exposure results in beneficial or adverse effects in
with soy, we observed mean total isoflavone concentrations of
infants remains to be determined. Except for allergies (41) and
2.0 ?mol/L in plasma, 0.2 ?mol/L in breast milk, and 3.0
reports of goiter disease (42), however, acute toxic effects or
?mol/h in urine.
chronic adverse effects in infants fed soy-based formulas have
Although direct evidence of the protective effects of genistein
not been observed.
and daidzein on cancer is derived exclusively from in vitro and
Although the amount of aglycones in soy-based formulas
recent animal studies (1, 43, 44), our findings in human milk
(9–13%) is relatively low compared with that of isoflavone con-
raise an important hypothesis. In addition to protecting mothers
jugates, the presence of aglycones is considerable in light of the
against breast and ovarian cancer, breast-feeding is known to be
effect of aglycones on the bioavailability of isoflavones in
beneficial to infants by contributing to the prevention of infec-
infants. In contrast with free aglycones, which are ready to be
tions, diabetes mellitus, multiple sclerosis (45), and sudden
absorbed after ingestion, acylated and nonacylated isoflavone
infant death syndrome (46) and by inducing better intellectual
glucosides may not be absorbed as efficiently by young infants
development (47). Our results add another significant item to the
because the intestinal flora is not completely developed to per-
growing list of benefits of breast-feeding. Cancer incidence and
form ?-glucosidic cleavage (10, 32). The impaired ability for
severity are significantly reduced when newborn or prepubertal
absorption may lead to decreased systemic isoflavone concentra-
animals are treated with only 3 doses of genistein (1, 44). The
tions. In contrast, isoflavones from breast milk may be readily
data presented here suggest that breast-feeding by a mother who
absorbed because of their presence as glucuronide and sulfate
consumes soyfoods may contribute to cancer prevention in
conjugates that can be hydrolyzed by endogenous enzymes.
infants because of exposure to daidzein and genistein, which are
Studies of the bioavailability of soy isoflavones in infants are
potential anticancer agents (2, 43). Although isoflavone doses
under way as part of efforts to further explore the potential
used in animal models are higher than those expected after soy
effects of these compounds in an early period of life.
exposure, a cancer-protective effect of isoflavones may take
In summary, our results show that the method of quantitating
place at an early and most critical developmental period and

---

1472S
FRANKE ET AL
TABLE 1
Iso?avone concentrations in 4 different soy-based infant formulas1
Prosobee
Alsoy
Gerbersoy
Isomil
Protein content (%)
17
17
17
14
Malonyldaidzin (?g/g)
29.5
23.6
27.1
19.4
Malonylglycitin (?g/g)
11.1
9.3
8.3
6.1
Malonylgenistin (?g/g)
48.5
38.7
47.2
35.4
Percentage malonylglucosides (%)
32
34
32
39
Acetyldaidzin (?g/g)
ND
ND
ND
ND
Acetylglycitin (?g/g)
ND
ND
ND
ND
Acetylgenistin (?g/g)
16.2
12.3
17.3
9.9
Percentage acetylglucosides (%)
6
7
7
6
Daidzin (?g/g)
53.1
21.3
48.5
22.7
Glycitin (?g/g)
12.9
4.7
11.0
6.7
Genistin (?g/g)
76.7
36.7
74.7
41.4
Percentage glucosides (%)
51
37
52
46
Daidzein (?g/g)
19.0
13.9
17.8
10.7
Glycitein (?g/g)
ND
ND
ND
ND
Genistein (?g/g)
14.4
7.5
9.3
2.8
Percentage aglucones (%)
12
13
10
9
Total daidzein
(?g/g)
101.6
58.7
93.4
52.7
(%)
36.1
35.0
35.8
34.0
Downloaded from
Total glycitein
(?g/g)
24.0
14.0
19.3
12.8
(%)
8.5
8.3
7.4
8.3
Total genistein
(?g/g)
155.7
95.2
148.6
89.5
(%)
55.3
56.7
56.9
57.7
www.ajcn.org
Total iso?avones (?g/g)
281.4
167.9
261.3
155.1
1 Concentrations are expressed as aglycone units determined by duplicate analysis. Mean CV for intraassay variability of analytes shown: 4.1% (range:
0.1–13.2%). Prosobee, Mead Johnson Company, Evansville, IN; Alsoy, Nestlé Company, Glendale, CA; Gerbersoy, Gerber Products Company, Fremont,
MI; Isomil, Ross Products Division of Abbott Laboratories, Columbus, OH. ND, not detected.
by on August 14, 2009
tion. J Agric Food Chem 1994;42:1674–7.
might protect an individual throughout life without requiring
6. Wang H, Murphy PA. Isoflavone content in commercial soybean
high doses. These findings may provide the basis for an alterna-
foods. J Agric Food Chem 1994;42:1666–73.
tive explanation of the lower cancer rates observed in Asian pop-
7. Tsukamoto C, Shimada S, Igita K, et al. Factors affecting isoflavone
ulations with high soya consumption. The lower cancer rates
content in soybean seeds: changes in isoflavones, saponins, and
observed in these populations may be due not to isoflavone expo-
composition of fatty acids at different temperatures during seed
sure by soy consumption in adulthood or childhood (48), but
development. J Agric Food Chem 1995;43:1184–92.
8. Coward L, Barnes NC, Setchell KDR, Barnes S. The antitumor
rather to exposure to isoflavones shortly after birth, in a critical
isoflavones, genistein and daidzein, in soybean foods of American
period of life, through the mother’s milk.
and Asian diets. J Agric Food Chem 1993;41:1961–7.
9. Barnes S, Kirk M, Coward L. Isoflavones and their conjugates in soy
We thank the participant in the study for her continuous support during the
foods: extraction conditions and analysis by HPLC–mass spectrom-
intervention trial. We are also grateful for the excellent performance of GC-
etry. J Agric Food Chem 1994;42:2466–74.
MS-SIM measurements by Hans Geyer, German Sports University, Cologne,
10. Setchell KDR, Welch MB, Lim CK. HPLC analysis of phytoestro-
Germany.
gens in soy protein preparations with ultraviolet, electrochemical,
and thermospray mass spectrometric detection. J Chromatogr
REFERENCES
1987;386:315–23.
11. Adlercreutz H, Mazur W. Phytoestrogens and Western diseases. Ann
1. Lamartiniere CA, Moore J, Holland M, Barnes S. Neonatal genis-
Med 1997;29:95–120.
tein chemoprevents mammary cancer. Proc Soc Exp Biol Med
12. Adlercreutz H, Fotsis T, Lampe J, et al. Quantitative determination
1995;208:120–3.
of lignans and isoflavonoids in plasma of omnivorous and vegetarian
2. Franke AA, Custer LJ, Mordan LJ, Cooney RV, Harwood PJ, Tanaka
women by isotope dilution gas chromatography–mass spectrometry.
Y. Inhibition of neoplastic transformation and bioavailability of
Scand J Clin Lab Invest 1993;53:5–18.
dietary phenolic agents. In: Manthey J, Buslig B, eds. Flavonoids in
13. Cassidy A, Bingham S, Setchell KD. Biological effects of a diet of
the living system. New York: Plenum Press, 1998.
soy protein rich in isoflavones on the menstrual cycle of pre-
3. Franke AA, Custer LJ, Cerna CM, Narala KK. Quantitation of phy-
menopausal women. Am J Clin Nutr 1994;60:333–40.
toestrogens in legumes by HPLC. J Agric Food Chem 1994;
14. Adlercreutz H, Fotsis T, Bannwart C, Wahala K, Brunow G, Hase T.
42:1905–13.
Isotope dilution gas chromatographic–mass spectrometric method
4. Kudou S, Fluery Y, Welti D, et al. Malonyl isoflavone glycosides in
for the determination of lignans and isoflavonoids in human urine,
soybean seeds (glycine max MERRILL). Agric Biol Chem 1991;
including identification of genistein. Clin Chim Acta
55:2227–33.
1991;199:263–78.
5. Wang H, Murphy PA. Isoflavone composition of American and
15. Kelly GE, Nelson C, Waring MA, Joannou GE, Reeder AY. Metabo-
lites of dietary (soya) isoflavones in human urine. Clin Chim Acta
Japanese soybeans in Iowa: effects of variety, crop year, and loca-

---

ISOFLAVONES IN HUMAN BIOLOGICAL FLUIDS
1473S
1993;223:9–22.
30. Lawrence RA. Breastfeeding. 4th ed. St Louis: Mosby, 1994.
16. Hutchins AM, Slavin JL, Lampe JW. Urinary isoflavonoid phytoe-
31. Hutchins AM, Lampe JW, Martini MC, Campbell DR, Slavin JL.
strogen and lignan excretion after consumption of fermented and
Vegetables, fruits, and legumes: effect on urinary isoflavonoid phy-
unfermented soy products. J Am Diet Assoc 1995;95:545–51.
toestrogen and lignan excretion. J Am Diet Assoc 1995;95:769–74.
17. Kurzer MS, Lampe JW, Martini MC, Adlercreutz H. Fecal lignan
32. Cruz MLA, Wong WW, Mimouni F, et al. Effects of infant nutrition
and isoflavonoid excretion in premenopausal women consuming
on cholesterol synthesis rates. Pediatr Res 1994;35:135–40.
flaxseed powder. Cancer Epidemiol Biomarkers Prev 1995;4:353–8.
33. Baird DD, Umbach DM, Lansdell L, et al. Dietary intervention
18. Adlercreutz H, Fotsis T, Kurzer MS, Wähälä K, Mäkelä T, Hase T.
study to assess estrogenicity of dietary soy among postmenopausal
Isotope dilution gas chromatographic–mass spectrometric method
women. J Clin Endocrinol Metab 1995;80:1685–90.
for the determination of unconjugated lignans and isoflavonoids in
34. Barnes S, Sfakianos J, Coward L, Kirk M. Soy isoflavonoids and
human feces, with preliminary results in omnivorous and vegetarian
cancer prevention: underlying biochemical and pharmacological
women. Anal Biochem 1995;225:101–8.
issues. In: American Institute for Cancer Research, ed. Dietary phy-
19. Franke AA, Custer LJ. High-performance liquid chromatography
tochemicals in cancer prevention and treatment. New York: Plenum
Press, 1996:87–100.
assay of isoflavonoids and coumestrol from human urine. J Chro-
35. Marzo A, Arrigoni Martelli E, Bruno G. Assay of hydroxyfarrerol in
matogr B Biomed Appl 1994;662:47–60.
biological fluids. J Chromatogr 1990;535:255–61.
20. Xu X, Wang H-J, Murphy PA, Cook L, Hendrich S. Daidzein is a
36. Wakui Y, Yanagisawa E, Ishibashi E, et al. Determination of baicalin
more bioavailable soymilk isoflavone than is genistein in adult
and baicalein in rat plasma by high-performance liquid chromatog-
women. J Nutr 1994;124:825–32.
raphy with electrochemical detection. J Chromatogr
21. Franke AA, Custer LJ, Wang W, Shi SJ. HPLC analysis of
1992;575:131–6.
isoflavonoids and other phenolic agents from foods and from human
37. Sfakianos J, Coward L, Kirk M, Barnes S. Intestinal uptake and bil-
fluids. Proc Soc Exp Biol Med 1998;217:263–73.
iary excretion of the isoflavone genistein in the rat: evidence for its
22. Coward L, Kirk M, Albin N, Barnes S. Analysis of plasma
enterohepatic circulation. J Nutr 1997;127:1260–8.
isoflavones by reversed-phase HPLC–multiple reaction ion moni-
38. Scheline RR. Handbook of mammalian metabolism of plant com-
toring–mass spectrometry. Clin Chim Acta 1996;247:121–42.
pounds. Boca Raton, FL: CRC Press, 1991.
Downloaded from
23. Franke AA, Custer LJ. Daidzein and genistein concentrations in
39. Griffith LA. Mammalian metabolism of flavonoids. In: Harborne
human milk after soy consumption. Clin Chem 1996;42:955–64.
JB, Mabry TJ, eds. The flavonoids: advances in research. London:
24. Geyer H, Schänzer W, Mareck-Engelcke U, Donike M. Factors
Chapman & Hall, 1982:681–718.
influencing the steroid profile. In: Donike M, Geyer H, Gotzmann
40. Eldridge AC, Kwolek WF. Soybean isoflavones: effect of environ-
G, Mareck-Engelke U, eds. Recent advances in doping analysis.
ment and variety on composition. J Agric Food Chem 1983;
Cologne, Germany: Sport und Buch Strauss Press, 1996:95–113.
31:394–6.
www.ajcn.org
25. Wang G, Kuan SS, Francis OJ, Ware GM, Carman AS. A simplified
41. Lalles JP, Peltre G. Biochemical features of grain legume allergens
HPLC method for the determination of phytoestrogen in soybean
in humans and animals. Nutr Rev 1996;54:101–7.
and its processed products. J Agric Food Chem 1990;38:185–90.
42. Shepard TH, Pyne GE, Kirschvink JF, Mclean M. Soybean goiter.
26. Lundh TJ-O, Pettersson H, Kiessling KH. Liquid chromatographic
N Engl J Med 1960;262:1099–103.
determination of the estrogens daidzein, formononetin, coumestrol,
43. Messina M, Persky V, Setchell KDR, Barnes S. Soy intake and can-
by on August 14, 2009
and equol in bovine blood plasma and urine. J Assoc Off Anal Chem
cer risk: a review of in vitro and in vivo data. Nutr Cancer 1994;21:
1988;71:938–41.
113–31.
27. Gamache PH, Maher TJ, Setchell KDR, Wu TH, Acworth IN. Sen-
44. Murrill WB, Brown NM, Zhang J-X, Manzolillo PA, Barnes S,
sitive analysis of phytoestrogens in animal tissue by HPLC with
Lamartiniere CA. Prepubertal genistein exposure suppresses mam-
coulometric array detection. Am J Clin Nutr 1998;68(suppl):1520S
mary cancer and enhances gland differentiation in rats. Carcinogen-
(abstr).
esis 1996;17:1451–7.
28. Bluck LJC, Bingham SA. Isoflavone content in breast milk and soy
45. Anonymous. A warm chain for breastfeeding. Lancet 1994;344:
formulas: benefits and risks. Clin Chem 1997;43:850–1.
1239–41 (editorial).
29. Markham KR. Techniques of flavonoid identification. New York:
46. Savage F. Breastfeeding and SIDS. MIDIRS Midwifery Digest
Academic Press, 1982.
1992;4:227–6.

---