Daidzein and genistein concentrations in human ... - Clinical Chemistry

3 downloads 0 Views 1MB Size Report
Daidzein and genistein concentrations in human milk after soy consumption. ADRIAN. A. FRANKE* and LAURIE J. CUSTER. Soy isoflavones were quantified.
‘Iinical Chemistry 55-964

42:6

(1996)

Daidzein and genistein concentrations in human milk after soy consumption ADRIAN

A.

FRANKE*

J.

and LAURIE

CUSTER

Soy isoflavones were quantified from human milk by a fast, precise, and selective HPLC method after enzymatic hydrolysis of conjugated isofiavones and extraction with ethyl acetate. Isoflavone aglycones and their mammalian metabolites equol and O-desmethylangolensin were separated selectively and identified by absorbance patterns, fluorometric and electrochemical detection, gas chromatography-

99.999%

mass spectrometric analysis after trimethylsilylation, and with internal and external authentic standards. HPLC injections of 20 pL of human milk showed detection limits of 1-3 pmol for all analytes by using diode-array detection. The detection limit could be improved by as much as 1000-fold by extended concentration through partitioning

on the target tissue or by imprinting mechanisms. Genistein and daidzein are also thought to play a major role in reducing cancer risk [1,14] because populations with high isoflavone exposure

with ethyl acetate, by increasing the injection techniques. We used isoflavone concentrations urine after challenge soybeans in the diet. potential of isoflavones fants exposed to these TERMS:

of genistein

approach for future breast

cancer

cells [12].This

most

striking

anticancer

was therefore suggested to be a promising clinical applications [12]. Additionally,

incidence

and tumor

numbers

were decreased

in

newborn mice given only three genistein doses [13]. This suggests a potent anticancer activity of this agent at a very early and critical period in life through direct effectsof the isoflavone

through soy consumption have low cancer rates [15, 16]. Human exposure to dietary genistein and daidzein occurs mainly through

using electrochemical detection, by volumes, or by combining these the proposed method to monitor in human milk and in human with 5, 10, and 20 g of roasted Implications of the results for the to prevent cancer in newborn inagents are discussed.

isoflavonoids . phytoestrogens O-desmethylangolensin #{149} breast-feeding . cancer newborns INDEXING

potency

of cancer

intake of soy food [17]. Concentrations

of these

agents

in soy average 2 g/kg (dry wt.) but vary greatly [18-21]. Traditionally, gas chromatography-mass spectrometry (GC-

MS) has been applied

to determine

soy isoflavones

and their

metabolites in human biological fluids, including urine [8, 22, 23], plasma [24], and feces [25, 26].’ Recently, the introduction of HPLC to measure these anaiytes in human urine [27, 28] has allowed the measurement of a varietyof phytoestrogens, including aglycones and conjugated analytes, in one run. Compared

#{149} equol

with GC-MS,

HPLC

required fewer steps for sample

preparation and analysis and demanded less technician time and less expensive instrumentation. In independent studies of soy intervention, HPLC [27] proved to be as accurate as GC-MS

prevention

Isoflavonoid compounds, specifically genistein and daidzein, have been implicated in the prevention of cancers [I], possibly through multiple effects connected with the inhibition of carci-

[8, 23] for measuring urinary isoflavone concentrations, as evidenced by similar results achieved in these trials. Although isoflavone metabolites such as equol and its deriv-

nogenesis [1-3]. Recent reports about activities of isoflavonoids-radical scavenging, antioxidant [4], antiestrogenic (5-8], antimutagenic, antiproliferative [9], differentiation-inducing [10], and angiogenesis-inhibiting [11] activities-add to the growing list of anticancer effects of these agents. Most importantly, genistein linked to an antibody against B cells led to 100% long-term survival in leukemic mice by killing

atives have been identified in cow’s milk [29, 30], to our knowledge the presence of dietary isoflavones in human milk has not been reported. In support of future studies favoring noninvasive protocols and assessing the potential cancer protective role of a diet containing

soya or isoflavones,

we developed

dure to determine, for the first isofiavones in human miik.

Cancer Research Center of Hawaii, 1236 Lauhala St., Honolulu, HI 96813. Author for correspondence. Fax 808-586-2973; e-mail adrian@crch. hawaii.edu. Received November 20, 1995; accepted February 8, 1996.

Nonstandard

abbreviations;

chromatography-mass

955

spectrometry;

SIM, and

time,

an HPLC

concentrations

proceof soy

CC-MS. SPE,solid-phaseextraction.

selected-ionmonitoring;

gas

956

Franke

Materials PARTICIPANTS

AND

and Custer:

Daidzein

and Methods

SAMPLES

One Caucasian and one Chinese woman between ages 32 and 34 years, both in week 15 postpartum, participated in this study. Both were in good health, of normal height and weight (55 and 57 kg), nonsmoking, not on any medication (including hormones or dietary supplements), and without particular dietary (e.g., not vegetarians). During the study, the participants did not consume any alcohol and maintained their usual diet except for the Caucasian woman’s additional intake of 5, 10, and 20 g of roasted soybeans at times 0, 24, and 72 h, respectively, during intervention. All procedures of the protocol followed were in accordance with the ethical standards of the Helsinki Declaration of 1975, as revised in 1983. Milk was collected from these women during breastfeeding patterns

from the breast not used in the feeding, starting first soybean intake. Milk samples were stored

12 h before the in plastic vials

between 6 #{176}C and -4 #{176}C. Isoflavones in milk were found to be stable for at least 5 days when kept in this temperature range

and genistein

in human

milk

kU/L) and arylsulfatase isolated from H. pomatia (1-5 kUIL) were purchased from Boehringer Mannheim (Indianapolis, IN). Equol and O-desmethylangolensin were purchased from K. W#{228}h#{228}l#{228}, University of Helsinki, Finland.

PROCEDURES

Extraction and acid hydrolysis of isoflavones from soybeans. Roasted soybeans were extracted as described previously [17]. In brief, 1 g of soybeans was homogenized and extracted with simultaneous hydrolysis by refluxing for I h in a mixture of 10 mL of conc. hydrochloric acid and 40 mL of 96% aqueous ethanol containing 20 Mg/L flavone (internal standard) and 0.25 g/L butylated hydroxytoluene. Enzymatic hydrolysis and extraction of isoflavones from human milk. We mixed 2-4 mL of human milk equilibrated to room temperature with 25 ML of flavone ethanol), 50 ML of 13-glucuronidase

solution (120 MgfL in 96% reagent (200 kU/L), and 50

ML of arylsulfatase reagent (5 kU/L) and then stirred this for 1 (data not shown). Six overnight urine samples and two additional h at 37 #{176}C. This sample was extracted threetimes with 2 mL of samples, 12 h before and 87 h after first soybean intake, were ethyl acetate (ACS-certified) and the organic phases were comstored in disposable bottles containing 0.2 g of boric acid and 0.1 bined after centrifugation; the combined phases were then dried g of sodium ascorbate to control for bacterial contamination and under nitrogen. The dry extract was redissolved in 150 ML of degradation of analytes. After mixing and determining the methanol by vortex-mixing, after which 50 ML of 0.2 mol/L volume of each urine sample, 50 mL was transferred into acetate buffer (pH 4) was added. After centrifugation, 150 ML of disposable plastic tubes and stored between -4 #{176}C and -20 #{176}C. clear sample was put in an insert of an amber vial for automated All collection times, including times of previous urine voids, HPLC injections of 20-ML samples. were recorded for adjustment purposes. In a parallel experiment, we mixed 25 ML of flavone (120 APPARATUS HPLC analyses were carried out on a “system Gold” chromatograph with a Model 507 autosampler and a Model 168 dualchannel diode-array detector (all from Beckman, Fullerton, CA), a Model FD 100 fluorescence detector (GTllSpectroVision, Concord, MA), and a Coulochem 11-5200 electrochemical detector (ESA, Bedford, MA) with a 5011 coulometric cell. Absorbance readings were obtained with a DU-62 spectrophotometer (Beckman). Samples were evaporated with a Savant AS 160 Speed-Vac (Savant, Farmingdale, NY) at room temperature. GC-MS analysis was carried out with a Hewlett-Packard (Palo Alto, CA) Model 5890 gas chromatograph using a Model 597 1A mass-selective detector and electron impact ionization at 70 eV.

REAGENTS

Methanol, hydrochloric acid, acetic acid, 96% ethanol, dimethyl sulfoxide, ethyl acetate, and all solvents used for HPLC and absorbance readings were analytical-grade or HPLC-grade from Fisher Scientific (Fair Lawn, NJ). Butylated hydroxytoluene, sodium acetate, genistin, and glucuronidase/sulfatase (isolated from Helix pomatia type HP-2S) were purchased from Sigma Chemical Co. (St. Louis, MO). Daidzein and genistein were obtained from ICN (Costa Mesa, CA), flavone from Aldrich (Milwaukee, Wi), NY), and 13-glucuronidase

coumestrol from Serva (New York, isolated from Escherichia co/i (200

Mg/L in 96% ethanol) with 130 ML of methanol and 45 ML of 0.2 mol/L acetate buffer (pH 4) in the same batch for calculating the recovery of the internal standard. Extraction and enzymatic hydrolysis of urinary isoflavones. Urine was extracted as described previously [27]. In brief, 20 mL of clear urine was mixed with 5.0 mL of 0.2 mol/L acetate buffer (pH 4) and 200 ML of flavone internal standard (60 Mg/L in 96% ethanol) and filtered through a preconditioned C18 reversedphase solid-phase extraction (SPE) column (PGC Scientific, Gaithersburg, MD). We then washed the column with 2 mL of acetate buffer and eluted the analytes with 100% methanol. The eluate was dried with the Speed-Vac at room temperature, redissolved in 1.0 mL of 0.2 mol/L phosphate buffer (pH 7.0), mixed thoroughly with 50 ML of 13-glucuronidase [31]and 50 ML of arylsulfatase, and incubated for 1 h at 37 #{176}C. Subsequently, the enzymes of the hydrolyzed samples were inactivated by addition of 0.9 mL of 100% methanol. Samples were analyzed immediately or stored at -20 #{176}C and analyzed by HPLC (after equilibration to room temperature, vortex-mixing, and centrifugation at 8SOg for 5 mm). The samples could be concentrated further by partitioning the isoflavones from the hydrolyzed sample into ethyl acetate. The combined organic phases were dried under nitrogen and redissolved in 150 ML of mobile phase plus 50 ML of 0.2 mol/L acetate buffer (pH 4); 20 ML of this was injected into the HPLC In a parallel experiment, we mixed

system. 200 ML of flavone

(60

Clinical Chemistry

42, No. 6, 1996

Mg/L in 96% ethanol) with 0.9 mL of buffer and 0.9 mL of methanol in the same batch to calculate recovery of the internal standard.

and previous

Trimethylsilylation. Dry milk extracts or crystalline authentic standards were dissolved in 0.1 mL of 20 g/L N-methyl-N(trimethylsilyl)trifluoroacetamide (MSTFA, Sigma) in imidazole and incubated for 15 mm at 60 #{176}C before GC-MS analysis.

HPLC

957

void (hours),

ery of the internal

standard

urine volume

(milliliters),

and recov-

(%).

Results and Discussion

Chromatography. All HPLC analyses were carried out on a 150 X 3.9 mm (i.d.) NovaPak C18 (4-Mm particles) reversed-phase column (Waters, Milford, MA) coupled to a 10 x 4.6 mm (i.d.) Adsorbosphere C18 (5-Mm particles) direct-connect guard column (Ailtech; Deerfield, IL). Elution was performed at a flow rate of 0.8 mL/min with the following linear gradient of solvent B in solvent A [B = methanol/acetonitrile/dichloromethane (10/5/1 by vol), A = acetic acid/water (10/90 by vol)]: 5% for 5 mm, from 5% to 45% in 20 mm, from 45% to 70% in 6 mm, and from 70% to 5% in 3 nun, with equilibration for 15 mm before subsequent injection. Analytes were monitored with the dual-channel diode-array detector at 260 and 280 nm and simultaneously with the coulometric detector at +500 mV during the entire HPLC run. Observed peaks were then scanned between 190 and 400 nm. Gas chromatography was performed by injecting 3 ML of trimethylsilylated sample onto a 17 m X 0.2 mm (i.d.) HewlettPackard Ultra-i capillary column (film thickness 0.11 Mm) and using helium as the carrier gas at a flow rate of 1.0 mL/min with a 1:10 split. We used the following temperature program: initial temperature 180 #{176}C, increasing at 10 #{176}C/min until reaching a final temperature of 320 #{176}C. Signals were registered in the selected-ion monitoring (SIM) mode at the following m/z ratios (determined after analysis of standards): daidzein = m/z 398, 383; genistein = m/z 471, 399, 228; equol = m/z 386, 192; O-desmethylangolensmn = m/z 459, 281. CALIBRATOR

SOLUTIONS

AND

HPLC

CALIBRATION

CONDITIONS,

DETECTION

LIMITS,

AND

CALIBRATION

HPLC separations of isoflavone aglycones and conjugates (Fig. I) were improved from our previous report [27] by including methanol and dichloromethane as additional organic modifiers in the mobile phase and by using a less-steep linear gradient. Better selectivity was achieved with the present system because it provided baseline separation of the soy isoflavones daidzein, glycitein, and genistein and of their acetyl and malonyl esters present in soy foods (not shown). In addition, the mammalian isoflavone metabolites equol and O-desmethylangolensin (Fig. 2) as well as other dietary phytoestrogens such as formononetin, biochanin-A, and coumestrol (Fig. 2E) were selectively separated. However, the present system led to laterretention times and thereby, to lower peak heights. Consequently, we calculated the detection limits for the analytes (Table 1) by using peak heights and found these to be slightly greater than those in the HPLC system used in our earlier report [27]. However, detection limits of analytes isolated from human milk and urine were 10-fold lower than those given in Table 1 because of the concentration step during extraction (seeMaterialsand Methods).Also, using higher volumes of starting material or increasing the injection volume could further lower the detection limits. More importantly, coulometric detection at + 500 mV improved (lowered)

OjJL0 DAIDThP4

GENISTEIN

FORMONONEFIN

BIOCHANIN.A

CURVES

Stock solutions of calibrators were prepared by dissolving 1-3 mg of the crystalline compound in 20 ML of dimethyl sulfoxide and then adding methanol to give concentrations of 2-5 molJL. Compounds with 0.996. Recovery of analytes from milk and urine was calculated from peak areas obtained after HPLC analyses according to the slopes of the calibration curves obtained from serial dilutions of the stock calibration solutions. Milk concentrations were expressed as nanomoles per liter after adjustment for recovery of the internal standard. Urine excretion was expressed as nanomoles per hour after adjusting for the time between urine collection

EL ANGOLENSIN

OUMFSROL

FLAVONE (internal standard)

Fig.1.Molecularstructure ofanalytesand internal standard(flavone).

958

Franke

and Custer:

Daidzein

and genistein

in human

milk

Table 1. HPLC detection limit and sensitivity. ECDb

uva detection limit, nmoi/L’

Daidzein Genistein Equol Q.Desmethylangolensin

54.3 26.6 164.2e 50.2e

Detection limit, Decrease of nmol/L#{176}detection limit”

15.8 13.9 29.7 85.2

Increase of sensitivity”

3.43 1.91 5.53 0.59 nd

3.44 1.94 7.90 1.01 nd

67.4 nd Cournestrol a Ultravioletabsorbanceat 260 nm. b Electrochemical values obtained by coulometric detection at 500 my. Determined by peak height with a 20-L HPLC injection atsignal-to-noise ratio = 5.

Coulometric values relative to absorbance values. Ultraviolet values obtained by absorbance at 280 nm. nd = not determined. d

#{176}

The present system also produced calibration curves having extremely high linearity in the concentration range of interest forall analytes included in this assay (see Materials and Methods). ISOFLAVONE

CONCENTRATIONS

In soybeans.Total concentrations of daidzein, glycitein in roasted soybeans obtained after

genistein, extraction

and and

simultaneous acid hydrolysis were 830, 913, and 174 mg/kg, respectively-values in good agreement with those reported earlier [17]. Consequently, the dosage applied during dietary intervention in this study amounted to 0.08, 0.15, and 0.30 mg/kg for daidzein and 0.08, 0.17, and 0.33 mg/kg for genistein-amounts found to exhibit biological effects [3], especially when given in combination [32].

Fig. 2. HPLC chrornatograrns of extracts from human milk (A-D) and from authentic standards (E) shown on the same scale. Extract of a milk sample collected 18 h after challenge with 20 g of roasted soybeans monitored at 260 nm (A) and simultaneously at 280 nm (C). Same sample after supplementation with daidzein (1), genistein (2), equol (3), and Odesmethylangolensin (4) with detection at 260 nm (B) and 280 nm (D). Trace E shows the chromatogram with no number assigned). mm (not shown). Analyte respectively: daidzein 3.7,

of authentic standards, including coumestrol (peak Flavone. used as internal standard. elutes after 31 amounts in traces A (= C), B (= D), and E are, 17.7, and 15.6 pmol; genistein 5.4, 7.9, and 11.7 pmol; equol 0 (below detection limit). 45.7, and 5.4 pmol; O.desmethylangolensin 0 (below

detection

limit), 12.1, and 57.7 pmol.

the detection limits by more than fivefold relative to those obtained by monitoring absorbance at the absorbance maximum for each analyte (Table 1). Although retention times were altered,the peak areas obtained by monitoring at the respective absorbance maxima of the analytes remained unchanged. Therefore, the present I-IPLC system showed sensitivity equal to that of the previously applied system [27], as evidenced by the nearly identical slopes for the calibration curves. Sensitivity was further improved by using electrochemical detection instead of monitoring ultraviolet absorbance (Table i)

Extractedfrom human urine.Because isoflavonoids occur in human urine mainly as sulfates and glucuronides [22, 33], hydrolysis is recommended before HPLC analysis so that a less-complex chromatogram is obtained [27]. The sequence of enzymatic hydrolysis and SPE did not change the final results, which suggests a lack of enzyme inhibitors in urine. Also, phase separation with ethyl acetate gave similar yields relative to SPE. However, performing SPE first was found to be the fastest and most convenient method and therefore, was utilized in this study, as previously described [27]. Hydrolysis with a glucuronidase/sulfatase mixture prepared from H. pomatia showed interfering peaks with isoflavone signals in the HPLC chromatogram of human milk extracts; therefore, we used in all assays glucuronidase from E. co/i,which lacked these interfering compounds. Extracted from human milk.Milk presents a unique challenge for isoflavone extraction because of the high content of proteins that easily form gel aggregates, leading to low recovery of analytes. Therefore, proteins have been removed from milk by precipitation [34,35] or filtration [36], or analytes have been extracted directly without affecting proteins [37]. Tests with isoflavonoid-supplemented milk samples revealed that SPE did

Clinical Chemisty

959

42, No. 6, 1996

Table 2. Recovery of soy isoflavones and their main metabolites added to aqueous solution or to human milk. O.Desmethyl-

Daidzein

Sample Water + enzyme (n = 4) Amount added, nmol Mean recovery, % CV, % Human milk (n = 2 each)

Genlsteln

angolensin

Equol

0.3-39.4 99.6 3.9

0.5-63.2

103.6 9.4

BB-0 63.2 96.0

39.4 93.2

63.2

39.4

93.1

93.8

16.7 209.5 93.1

20.3 285.1 76.1

11.4 209.5 82.3

19.1 285.1

65.8 1833.1 127.9 98.5

37.9 765.6 93.8 87.9

Amount added, nmol Mean recovery, % Su-1 Amount added, nmol (n = 2) Mean recovery, % 886-3

Initial amount, pmol Amount added, pmol Mean recovery, % BB7-1 initial amount, pmol Amount added, pmol Mean recovery, % BB9-1 Initial amount, pmol Amount added, pmol Mean recovery, % Overall mean recovery in human milk, % n

=

number of samples tested; nd

=

not detected;

+ =

82.7

present, but belowthe limit

not separate lipids and proteins from the aqueous phase and that many solvents were not applicable for isoflavone extractions. Applying hexane for lipid removal led to precipitates, dichloromethane extracted mainly lipids, and diethyl ether formed gels, preventing isolation of the analytes. In contrast, ethyl acetate, which is frequently used for isolation of dietary phenolic [38] and flavonoid [39-41] compounds, resulted in selective concentration of the analytes in the organic phase and led to greater recoveries than were obtained with tert-butylmethyl ether. Thus recoveries of isoflavone aglycones from human milk

+

+

1491.5 69.8

698.3 71.1 + 698.3 85.2

+

1491.5 81.5 +

+

9689.2 111.5 89.3

37354.4 108.8 86.7

of quantification.

were high (Table 2), and glucosidic conjugates were not coextracted (data not shown). Therefore, hydrolysis of isoflavone conjugates was required before extraction so as to include these conjugates in the assay. The lack of detectable isoflavone concentrations after extraction without hydrolysis suggested that all isoflavones in human milk are present as conjugates. Precision studies and analytical recoveries of analytes added to milk samples (Table 2) were found to be within accepted limits for phytoestrogen analyses [22-24/ and so confirmed the validity of the proposed procedure.

B

A 4 .0 IC .0

Fig. 3. Absorbance

scans of genistein

(A)

and daidzein (B). Absorbance scans for authentic standards (two upper traces in A and two lower traces in B) are with those for material obtained in HPLCpeaks 2 (lower trace in A) and I (upper trace in B) from human milk extracts. identical

nm

nm

960

Franke

and Custer:

Daidzein

and genistein

in human

lan 398.00

milk

amu from

36010404

25000 20000 15000

101)00

8.1

8.2

0.4

0.3

Scan 509 (0.387

B

0.5

0.7

0.6

0.8

0.9

9.0

Thue(mm.) mUi)of 36010484 SUBTRACTED 390

20000

303

150011 10000

-

5000

201

192

0 150

200

250

300

350

400

Masge

Ion 398.00 aniu from 36010014

AmdoncaC

03

0.6

8.8

9.0

TIme (mm.) Scan 482 (0.369 mm) 0(36010014 SUBTRACTED

D

500000

398

303

400000

0

300000 200000

Fig. 4. Gas chromatograms (A, C) and fragmentation patterns (B, D) of trimethylsilylated milk extracts (A, B) and authentic daidzein (C, D) obtained by GC-MS/SIM analysis.

IDENTIFICATION HUMAN

MILK

OF ISOFLAVONOIDS AND

DAIDZEIN

100000

220

192 II

TMS

ETHER

201

,,t,i,,i,i,u,#{149},i,

150

200

I

250

300

I

350

400

Mass/tharge

EXTRACTED

lowed by GC-MS analysis in SIM mode again indicated presence of daidzein and genistein in human milk because

FROM

URINE

the the

Analytes

(Fig. 1) were routinely identified by retention times in various HPLC systems and by ultraviolet absorption patterns obtained by diode-array detection (Fig. 3) in comparison with

GC retention times and the mass fragmentation patterns were identical to those of authentic standards (Fig. 4 and 5). These GC-MS values were also in excellent agreement with those

authentic standards and with reported absorbance data [42, 43]. In addition, we used fluorometric [27, 41] and coulometric [44]

reported previously [26]. None of the other soy isoflavones or metabolites shown in Fig. 1 was detected in human milk extracts with this GC-MS/SIM method, from which we deduce their

detection milk.

to confirm

Trimethylsilylation

the presence

of these

of representative

agents milk

in human extracts

fol-

absence

in human

milk after soy consumption.

961

Clinical Cheinistiy42, No. 6, 1996

Ion 471.00ann. front 36010404

A 10000 8000

: 2000

8.4



I

‘8

‘8.5’

111111111

I

0.9

I

9.)

I

9.1

I

92

I

TIme (.)

B

of 36010404

Scan 544 (0.686)

SUBTRACTED 471

8001 7001

6801 5801

4001

228

3001) 2001 1001 II

-

192

201

I I

I’



I#{149}

I

I

I

I

-r---

,

I

200

400

Masthae

-

dancaC 5000001 4000001 6800001

,I

I

I

I

I

3000 W00 1000001

0

III

0.4

8.5

J)

Pxmdance

0.7

0.6

)

Scan 530(0.694

0.8 ‘ TIme (mm.)

0.9

9.09.192

I

I

0(36010014 SUBTRACTED

#{149}#{231}J(.1_ .

471

6000001 5000001 4000001

SMTO

Fig. 5. Gas chromatograms

(A, C) and

(B, D) of trimethylsilylated milk extracts (A, B) and authentic genistein (C, 0) obtained by GC-MS/ SIM analysis. fragmentation

ISOFLAVONE AFTER

SOY

patterns

CONCENTRATIONS

IN HUMAN

0

GFNISTEINThIS ETHER

3000001

228

2000001

192

1000001

0

,

281 \;.

I

I

I

383

(

399

I

-r

,

200

300

400

Mas,!thgs

MILK

AND

URINE

CHALLENGE

Urine analysis by this procedure revealed patterns (Fig. 6) identical to those of the conjugated soy isoflavones daidzein, genistein, and glycitein and of their main metabolites equol and O-desmethylangolensin, as reported previously [23, 27]. In contrast, analysis of human milk showed that only daidzein and genistein conjugates were above the HPLC detection limit, even after challenge with 25 g of roasted soybeans (data not shown). The exclusive occurrence of the major soy isoflavones in milk is

not surprising; glycitein exposure through soy challenge is minor due to its low concentrations in soybeans [23, 45]. Preferential excretion of the metabolites over the parent isoflavones is suggested by the isoflavone:metabolite ratio, which is seen to be much higher in plasma than in urine or feces [24, 26]. Because milk is produced by secretory processes of blood [46],the low plasma concentrations of isoflavonoid metabolites in human blood might explain their absence in milk. Soybean challenge led to a fast and dose-dependent response

962

Franke

and Custer:

Daidzein

and genistein

in human

milk

z z,-’

0

c)

z

0

Fig. 6. Concentrations of daidzein and genistein

in human milk (A) and urine (B) after challenge with 5, 10, and 20 g of roasted soybeans as determined by diode-array I-IPLC. Milk samples were collected at the time of feeding, starting 12 h before first soybean intake. Six overnight urine samples and two

r

additional samples, 12 h before and 87 h after first soybean intake, were obtained to monitor urinary isoflavone excretion.

in human milk (Fig. 6). Maximum milk concentrations were reached 10-14 h after soy intake, and baseline was reached 2-4 days later, depending on the dose. The isoflavone patterns in milk followed those in urine except with a slight delay (Fig. 6). This is in good agreement with many other polar micronutrients or drugs, which show a faster urinary excretion than secretion into milk [46]. Most importantly, however, the Caucasian woman showed higher concentrations of genistein conjugates than those of daidzein conjugates in plasma and, because of secretory processes, also in milk. This is in contrast to higher daidzein than genistein concentrations observed in breast milk of a Chinese woman (see below), a pattern generally reported from plasma [24], urine [22, 23, 27, 28], and feces [26], which suggests large interindividual variation. Milk concentrations of daidzein and genistein conjugates increased rapidly aftersoy consumption; this was followed by their rapid decrease and, most interestingly, by a subsequent increase before reaching baseline values (Fig. 6). This biphasic elimination pattern has already been observed in animal plasma

5gsoy

12

logsoy

36

2ogsoy

Time

(h)

after treatment with the flavane hydroxyfarrerol [47] or the flavone baicalin [39] and, more importantly, in human plasma [33] and urine [27] aftersoy intervention. This biphasic phenomenon has been suggested to result from enterohepatic circulation [33],a process known to occur with flavonoids [48]. Enterohepatic circulation might also, therefore, explain the biphasic pattern observed in human milk. Isoflavone concentrations in three milk and urine samples from a Chinese woman eating her usual diet (including tofu soup once a day for dinner) were analyzed by the proposed procedure. The concentrations of daidzein and genistein in her milk (80-110 and 30-50 nmol/L, respectively) were similar to those observed in the Caucasian woman’s milk after challenge with roasted soybeans. This is in good agreement considering the similar total isoflavone exposure from these two food items [17]. However, urinary excretion of these compounds was less by the Chinese woman (80-150 nmollh daidzein and 8-33 nmollh genistein) than by the Caucasian woman during soybean intervention. This difference may reflect interindividual variation in

963

Clinical Chemistry 42, No. 6, 1996

excretion [23, 27] or the women’s ingestion of different food items. Although demonstrated exclusively in carcinogenor cytokine-induced cell and animal models, the anticancer properties of the soy isoflavones genistein and daidzein are particularly intriguing when considered in light of these results. Breastfeeding is known to be beneficial not only to the mother, by limiting fertility and protecting against ovarian and breast cancer, but also particularly to the infant. Newborn babies are protected by mother’s milk from a variety of diseases e.g., various types of infections, diabetes mellitus, and multiple sclerosis [49].Additionally, lower incidence of sudden infant death syndrome [50] and better intellectual development [51] are attributed to breast-feeding. Our results add another important item to the growing list of benefits of breast-feeding, a decrease in cancer incidence and severityboth of which are significantly reduced when newborn animals are treated with only three single doses of genistein [13].The data presented here suggest a cancer-preventive effect of breast-feeding to the offspring when mothers consume soy foods, in that such infants would be exposed to the known anticancer agent genistein and also daidzein. This effect, which would take place at a very early and most critical developmental period, might protect the individual throughout life. Also, our findings may provide the basis for an alternative explanation for the lower cancer rates observed in Asian populations with high consumption of soya. The lower cancer rates in these populations might not be the result of isoflavone exposure by soy consumption in adulthood or childhood [52] but rather of isoflavone exposure shortly after birth, in a critical period of life, through mother’s milk containing these agents. Conceivably, the isoflavone conjugates obtained from mother’s milk are more bioavailable to the newborn child than are the conjugates from soy foods. Young infants might not be able to absorb isoflavones from soya because their gut flora are incompletely developed, preventing hydrolysis of the acylated and nonacylated isoflavone glucosides present in soy foods. Biotransformation and bioavailability studies of soy isoflavones in infants are required to further explore the cancer-preventive effects of these compounds in humans.

antipromotional effects of the soybean isoflavone genistein. Proc Soc Exp Biol Med 1995:208:124-30. 5. Mousavi Y, Adlercreutz H. Genistein is an effective stimulator of sex hormone-binding globulin production in hepatocarcinoma human liver cancer cells and suppresses proliferation of these cells in culture. Steroids 1993;58:301-4. 6. Adlercreutz H, Bannwart C, W#{228}h#{224}l#{228} K, M#{228}kel#{228} I, Brunow G, Hase T, et al. Inhibition of human aromatase by mammalian lignans and isoflavonoid phytoestrogens. J Steroid Biochem Mol Biol 1993; 44:147-53.

7. Makela SI, Pylkkanen LH, Santti RSS, Adlercreutz H. Dietary soybean may be antiestrogenic in male mice. J Nutr 1995;125: 437-45. 8. Cassidy A, Bingham 5, Setchell KDR. Biological effect of a diet of soy protein rich in isoflavones on the menstrual cycle of premenopausal women. Am J Clin Nutr 1994:60:333-40. 9. Hirano I, Gotoh M, Oka K. Natural flavonoids and lignans are potent cytostatic agents against human leukemic HL-60 cells. Life Sci 1994;55:1061-9. A, Huberman E. Genistein as an inducer of tumor cell differentiation: possible mechanisms of action. Proc Soc Exp Biol Med 1995:208:109-15. 11. Fotsis 1, Pepper M, Adlercreutz H, Hase 1, Montesano R, Schweigerer L. Genistein, a dietary ingested isoflavonoid, inhibits cell proliferation and in vitro angiogenesis. J Nutr 1995;125(Suppl):

10. Constantinou

790-7. Uckun FM, Evans E, Forsyth Ci. Waddick KG, Ahlgren LI, Chestrom LM, et al. Biotherapy of B-cell precursor leukemia by targeting genistein to CD19-associated tyrosine kinases. Science 1995:267:886-91. 13. Lamartiniere CA, Moore J, Holland M. Barnes S. Neonatal genistein chemoprevents mammary cancer. Proc Soc Exp Biol Med 1995;208:120-3. 14. Barnes S. Peterson 1G. Grubbs C, Setchell KDR. Diet and cancer: markers, prevention and treatment. New York: Plenum Press,

12.

1994:135-47. 15. Adlercreutz H, Honjo H, Higashi A, Fotsis T, Hamalainen E, Hasegawa T, et a). Urinary excretion of lignans and isoflavonoid

phytoestrogens in Japanese men and women consuming a traditional Japanese diet. Am I Clin Nutr 1991:54:1093-100. 16. AdlercreutzH, Markkanen H, Watanabe S. Plasmaconcentrations of phyto-oestrogensin Japanese men. Lancet 1993:342:120910.

17. Franke AA, Custer phytoestrogens

U,

Cema

in legumes

by

CM, Narala KK. Quantitation of HPLC. J Agric Food Chem 1994:42:

1905-13. Wang H, Murphy PA. Isoflavone composition of American and Japanese soybeans in Iowa: effects of variety, crop year, and location. J Agric Food Chem 1994:42:1674-7. 19. Tsukamoto C, Shimada 5, Igita K,Kudou 5, Kokubun M, Okubo K, et al. Factors affecting isoflavone content in soybean seeds: changes in isoflavones, saponins, and composition of fatty acids at different temperatures during seed development. I Agric Food Chem 1995:43:1184-92. 20. Barnes 5, Kirk M, Coward L. Isoflavones and their conjugates in soy foods: extraction conditions and analysis by HPLC-mass spectrometry. I Agric Food Chem 1994;42:2466-74. 21. Coward L, Barnes NC, Setchell KDR, Barnes S. The antitumor isoflavones, genistein and daidzein, in soybean foods of American 18.

We

are

very

grateful

for the most

generous

support

of the

participants and for the skillful performance of the GC-MS measurements by Hans Geyer, German Sports University, K#{246}ln, Germany.

1.

Messina

M, Persky

cancer risk: a review 1994;21:113-31.

References V, Setchell KDR, Barnes S. Soy intake and of in vitro and

in vivo data.

Nutr Cancer

2. Messina M, Barnes S. The role of soy products in reducing risk of cancer. J NatI Cancer Inst 1991;83:541-6. 3. Barnes 5, Peterson 1G. Biochemical targets of the isoflavone genistein in tumor cell lines. Proc Soc Exp Biol Med 1995:208: 103-8. 4. Wei H, Bowen R, Cai Q, Barnes S, Wang Y. Antioxidant and

22.

and Asian diets. I Agric Food Chem 1993;41:1961-7. Adlercreutz H, Fotsis I, Bannwart C, W#{228}h#{228}l#{228} K, Brunow G, Hase I. Isotope dilution gas chromatographic-mass spectrometric method for the determination of lignans and isoflavonoids in human urine, including identification of genistein. Clin Chim Acta

1991:199:263-78.

964

Franke

and Custer:

Daidzein

23. Kelly GE, Nelson C, Waring MA, Joannou GE, Reeder AY. Metabolites of dietary (soya) isoflavones in human urine. Clin Chim Acta 1993;223:9-22. 24. Adlercreutz H, Fotsis 1, Lampe I, W#{228}h#{228}l#{228} I, M#{228}kel#{228} I, Brunow G, et al. Quantitative determination of lignans and isoflavonoids in plasma of omnivorous and vegetarian women by isotope dilution gas chromatography-mass spectrometry. Scand I Clin Lab Invest 1993;53:5-18. 25. Kurzer MS. Lampe 1W, Martini MC, Adlercreutz H. Fecal lignan and isoflavonoid excretion in premenopausal women consuming flaxseed powder. Cancer Epidemiol Biomarkers Prey 1995;4:353-8. 26. Adlercreutz H, Fotsis I, Kurzer MS, W#{228}h#{228}l#{228} K, M#{228}kel#{228} 1, Hase 1. Isotope dilution gas chromatographic-mass spectrometric method

for the determination of unconjugated lignans and isoflavonoids in human feces, with preliminary results in omnivorous

and vegetarian women. Anal Biochem 1995:225:101-8. 27. Franke AA, Custer U. High-performance liquid chromatography assay of isoflavonoids and coumestrol from human urine. I Chromatogr B 1994;662:47-60. 28. Xu X, Wang H-I, Murphy PA, Cook L, Hendrich S. Daidzein isa more bioavailable soymilk isoflavone than is genistein in adult women. J Nutr 1994;124:825-32. 29. Bannwart C, Adlercreutz H, Fotsis I, W#{228}h#{228}l#{228} K, Hase I, Brunow G. Identification of isoflavonic phyto-oestrogens and of lignans in human urine and in cow milk by GC-MS. In: Todd JFJ, ed. Advances in mass spectrometry 1985. Proceedings of the 10th International Conference on Mass Spectrometry. Chichester, UK: John Wiley, 1986:661-2. 30. Bannwart C, Adlercreutz H, W#{228}h#{228}l#{228} K, Kotiaho I, Hesso A, Brunow G, et al. Identification of the phyto-oestrogen 3’ ,7-dihydroxyisoflavan,an isomerofequol, in human urine and cow’s milk. Biomed Environ Mass Spectrom 1988;17:1-6. 31. Donike M. Steroidprofiling in Cologne.In:Donike M, Geyer H, Gotzmann A, Mareck-Engelke U, Rauth 5, eds. 10th Cologne Workshop on Dope Analysis. K#{246}ln: Sport und Buch Strau6, 1993:1-5. 32. Franke AA, Mordan U, Cooney RV, Harwood PJ, Wang W. Dietary phenolic agents inhibit neoplastic transformation and trap toxic NO2 [Abstract]. 86th Meeting of American Association for Cancer Research,March 18-22, 1995, Toronto, Canada. Abstract #741. 33. Sfakianos I, Barnes S. The isoflavone genistein (GEN) undergoes an enterohepatic circulation. Hepatology 1996;23:in press. 34. Cunniff P. Method 932.05. In: Official methods of analysis of AOAC International, 16th ed. Arlington, VA: AOAC International, 1995. 35. Cunniff P. Method 991.21. In: Official methods of analysis of AOAC International, 16th ed. Arlington, VA: AOAC International, 1995.

and genistein

in human

milk

36. Cunniff P. Method 992.22, In: Official methods of analysis of AOAC International, 16th ed. Arlington, VA: AOAC International, 1995. 37. Cunniff P. Method 970.52. In: Official methods of analysis of AOAC International, 16th ed. Arlington, VA: AOAC International, 1995. 38. Kermasha 5, GoetghebeurM, Dumont I.Determination of phenolic compound profiles in maple products by high-performance liquid chromatography.

J Agric Food E, Ishibashi Sasaki H, et al. Determination of plasma by high-performance liquid

39. Wakui Y, Yanagisawa

Chem 1995;43:7O8-16. E, Matsuzaki Y, Takeda 5, baicalin and baicalein in rat chromatography with electro-

chemical detection. J Chromatogr 1992;575:131-6. 40. Lee M-J, Wang Z-Y, Li H, Chen L, Sun Y, Gobbo 5, et al. Analysis of plasma and urinary tea polyphenols in human subjects. Cancer Epidemiol Biomarkers Prey 1995;4:393-9. 41. Lundh TJ-0, Pettersson H, Kiessling KH. Liquid chromatographic determination of the estrogens daidzein, formononetin, coumestrol and equol in bovine blood plasma and urine. I Assoc Off Anal

Chem 1988:71:938-41. 42. Wang H, Murphy PA. Isoflavone content in commercial soybean foods. J Agric Food Chem 1994;42:1666-73. 43. Franke AA, Custer U, Cerna CM, Narala K. Rapid HPLC analysis of dietary phytoestrogens from legumes and from human urine. Proc Soc Exp Biol Med 1995:208:18-26. 44. Gamache P. Analysis of phenolics and flavonoids having potential biological activity using HPLC with coulometric array detection. NCI meeting on dietary phytoestrogens: cancer cause or prevention? Sept 21-23, 1994, Herndon, VA. Abstract #2. 45. Tsukamoto C,Shimada 5, Igita K,Kudou 5, Kokubun M, Okubo K, et al. Factors affecting isoflavone content in soybean seeds: changes in isoflavones, saponins, and composition of fatty acids at different temperatures during seed development. I Agric Food Chem 1995:43:1184-92. 46. Lawrence RA. Breastfeeding, 4th ed. St. Louis: CV Mosby, 1994: 323. 47. Marzo A, Arrigoni Martelli E, Bruno G. Assay of hydroxyfarrerol in biological fluids. I Chromatogr 1990;535:255-61. 48. Griffith LA. Mammalian metabolism of flavonoids. In: HarborneJB, Mabry TJ, eds. The flavonoids:

advances

in researchs.

London:

Chapman & HaIl, 1982:681-718. 49. A warm chain for breastfeeding [Editoriall. Lancet 1994:344: 1239-41. 50. Savage F. Breastfeeding and SIDS. Midwifery Dig1992;4:226-7. 51. Rogan WJ, Gladen BC. Breastfeeding and cognitive development. Early

Hum Dev 1993:31:181-93.

52. Saio K. Dietary pattern and soybean processingin Trop Agric Res Serv 1984;17:153-61.

Japan

today.