Chem. Res. Toxicol. 2001, 14, 1523-1528
1523
Detection of PhIP (2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine) in the Milk of Healthy Women Lillian S. DeBruin,† Perry A. Martos,‡ and P. David Josephy*,† Guelph-Waterloo Centre for Graduate Work in Chemistry (GWC2), Department of Chemistry & Biochemistry, Laboratory Services, University of Guelph, Guelph, Ontario N1G 2W1, Canada Received August 14, 2001
An increased risk of breast cancer has been observed in women who consume “very welldone” meats. Heterocyclic amines are mutagenic and carcinogenic pyrolysis products formed during high temperature cooking of meats. In the present study, human milk samples were analyzed for PhIP, one of the most abundant dietary heterocyclic amine. A protocol was developed with a mixed-mode cation exchange sorbent for the extraction of heterocyclic amines from milk. Milk samples were acquired from healthy Canadian women. With LC/MS analysis and the method of isotope dilution for quantification, levels of PhIP were determined in human milk samples. PhIP was detected in 9 of the 11 milk samples, at levels as high as 59 pg/mL (ppt). No PhIP was detected in the milk of the vegetarian donor. Detection of PhIP in milk indicates that ductal mammary epithelial cells are directly exposed to this carcinogen, suggesting that heterocyclic amines are possible human mammary carcinogens.
Introduction Diet plays a significant role in the development of human cancers. Many epidemiological studies have examined the role of diet in breast cancer, especially with regard to meat consumption. Increased risk of breast cancer among women with high dietary meat intake has been reported in multiple studies (1-3). Such a correlation may be due to one or more factors associated with high-meat diets, including fat intake (4), caloric intake (5), hormone exposure, and heterocyclic aromatic amines. Heterocyclic amines are pyrolysis products formed in muscle meats during grilling, broiling, or frying. These compounds are highly mutagenic in short-term assays and have been shown to be carcinogenic at many sites in animal models, notably the rat mammary gland (6). Heterocyclic amines undergo metabolic activation catalyzed by hepatic P4501A2 (or mammary P4501A1 and P4501B1) leading to formation of DNA adducts. Alternatively, these compounds may be bioactivated by peroxidases present in the mammary gland (7, 8). The total heterocyclic amine intake resulting from a typical Western diet is about 1.4 µg/day (9). The heterocyclic amines present at highest levels in the diet are PhIP,1 ARC, MeIQx, diMeIQx, and IQ, in approximate order of abundance. Several epidemiological studies have tested the hypothesis that dietary heterocyclic amine * To whom correspondence should be addressed. Phone: (519) 8244120 ext. 3833. Fax: (519) 766-1499. E-mail:
[email protected]. † Guelph-Waterloo Centre for Graduate Work in Chemistry (GWC2). ‡ Laboratory Services. 1 Abbreviations: ARC, 2-amino-R-carboline; diMeIQx, 2-amino-3,4,8trimethylimidazo[4,5-f]quinoxaline; Glu-P-2, 2-amino-dipyrido[1,2-a: 3N,2N-d]imidazole; harman, 1-methyl-9H-pyrido[4,3-b]indole; IQ, 2-amino-3-methylimidazo[4,5-f]quinoline; MeIQ, 2-amino-2,4-dimethylimidazo[4,5-f]quinoline; MeIQx, 2-amino-3,8-dimethylimidazo[4,5-f]quinoxaline; PhIP, 2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine; ppt, parts per trillion; RSD, relative standard deviation; SIR, selected ion recording.
exposure, as estimated by consumption of cooked meats, is a risk factor for breast cancer. In one such study, women who usually consumed “very well-done” red meat had a 4.6-fold increased risk of breast cancer compared to those who usually ate “rare” or “medium” done meat (10). The increased risk of breast cancer has been correlated specifically with exposure to PhIP but not to MeIQx or diMeIQx (11). Ames assay-positive mutagens have been reported in human milk (12, 13), nipple aspirate fluid (14), and mammary lipid (15). Although the compounds responsible for this activity have not been fully characterized, heterocyclic amines or polycyclic aromatic hydrocarbons, or their metabolites, are likely candidates. Analysis by 32 P postlabeling of DNA adducts in human mammary tissue also indicates the presence of hydrophobic aromatic adducts (16, 17). The analysis of trace chemical contaminants in human milk provides a valuable biomarker of exposure and indicates that the compounds have been biodistributed to the mammary tissue. In a previous study (18), aniline and derivatives (monocyclic aromatic amines) were detected in human milk. Excretion of dietary heterocyclic amines has been measured by analysis of human urine (19-22); however, analysis of these compounds in human milk has not yet been reported. Published methods for analysis of heterocyclic amines in foods and urine include the tandem solid-phase extraction procedure (23), the use of blue rayon (24), and liquid-liquid extraction (25). Here, we have developed a simple solid-phase extraction protocol utilizing a mixmode sorbent for the isolation of heterocyclic aromatic amines from milk. In this pilot study, human milk samples were analyzed for PhIP, the most abundant heterocyclic amine in typical diets. We report the presence of PhIP at parts per trillion (pg/mL) levels in many human milk samples. The exposure of the mammary
10.1021/tx015556u CCC: $20.00 © 2001 American Chemical Society Published on Web 10/16/2001
1524
Chem. Res. Toxicol., Vol. 14, No. 11, 2001
ductal epithelial cells, the cellular targets of breast tumorigenesis, to dietary heterocyclic amine mutagens may contribute to risk of breast cancer.
Experimental Procedures Chemicals and Materials. ARC, Glu-P-2, MeIQx, and PhIP were purchased from Toronto Research Chemicals, Inc. (Toronto, Ont., Canada). Harman was obtained from Sigma Chemicals (Sigma-Aldrich, Mississauga, ON, Canada). The stable isotope analogue of PhIP ([d5]-PhIP, labeled in the phenyl ring) was the kind gift of Dr. Robert Turesky (Nestec Ltd., Nestle Research Center, Lausanne, Switzerland; now at Nat. Center for Tox. Res., Jefferson, AR). Oasis MCX extraction cartridges (3 cm3/ 60 mg, 60 µm particle size) were purchased from Waters Limited (Mississauga, Ont., Canada). All solvents and reagents were of HPLC grade. Pooled bulk bovine milk was obtained from the Milk Testing Laboratory (University of Guelph, Canada) and frozen until needed. Caution: Heterocyclic amines are highly potent mutagens and suspect carcinogens and should be handled appropriately in a fume hood. Individual stock solutions were prepared to a concentration of 1 mg/mL in methanol, with the exception of Glu-P-2 HCl, which was prepared in 0.05 M HCl. For the MCX method development (see below), a mixed standard solution containing ARC, Glu-P-2, harman, MeIQx, and PhIP was prepared to a final concentration of 10 µg/mL in 10 mM ammonium acetate buffer, pH 4.0/acetonitrile (95:5). For the LC/ MS analysis of PhIP (see below), working solutions (0.1 and 1 µg/mL) were prepared in acetonitrile:water (50:50). The concentration of the [d5]-PhIP solutions was calculated from the molar extinction coefficient [19 400 at 316 nm (26)]. This deuterated labeled standard was stored in methanol with 10% 0.01 M acetic acid at -20 °C. Solutions for the generation of the calibration curve contained 1 ng/mL [d5]-PhIP and 0.5-100 ng/mL PhIP. Instrumentation and Chromatographic Conditions. Analyses for the method development of the MCX protocol were performed on a Shimadzu LC-10AD liquid chromatograph equipped with GT-104 degasser, SCL-10A system controller, CTO-10A column oven and a SPD-10AV UV-vis spectrophotometric detector in series with a RF-10A programmable spectrofluorometric detector. The system was controlled with CLASSVP software. The spectrophotometric detector was set to 263 nm for detection of MeIQx. Detection of the other heterocyclic amines was achieved as follows: at 0.0 min, λex ) 360 nm, λem ) 450 nm (for Glu-P-2); at 10.0 min, λex ) 300 nm, λem ) 440 nm (for harman); at 14.0 min, λex ) 315 nm, λem ) 390 nm (for PhIP); at 17.0 min, λex ) 335 nm, λem ) 410 nm (for ARC) (23). The heterocyclic amines were separated on a Phenomenex IBSIL 5 µm C18 BDS (4.6 mm × 250 mm) column (Torrance, CA) with a guard column (4.6 mm × 30 mm) of the same material. Flow rate was 1 mL/min throughout and the column oven temperature was set to 25 °C. Mobile phase A was 95% 10 mM ammonium acetate, pH 4.0/5% acetonitrile; mobile phase B was 95% acetonitrile/5% 10 mM ammonium acetate, pH 4.0. The gradient program began at 10% B and increased to 35% B by 12.0 min, 50% B by 15 min, and then to 100% B by 16 min, which was held for 4 min. For the LC/MS analysis, analytical separations were performed with a Hewlett-Packard Series 1090 liquid chromatograph equipped with a binary pump system. The column (Symmetry Shield RP18, 2.1 mm × 100 mm, 3.5 µm, and guard, 2.1 mm × 10 mm, 3.5 µm [Waters Limited (Mississauga, ON)] was connected with PEEK tubing (0.005 in. i.d.) to the electrospray interface of a VG Quattro II triple quadrupole mass spectrometer. The instrument was controlled with MassLynx (ver. 2.1) software and operated in the electrospray positive ion mode ionization (ESI+). The drying gas (N2) flow rate was set to 300 L/h with the nebulizing gas (N2) set to 15 L/h. The MS conditions for the optimum detection of PhIP were 3.5 kV capillary, 0.5 kV HV lens, 30 V cone, and 150 °C source
DeBruin et al. temperature. In selected ion recording (SIR) mode, the [M + H]+ ions of PhIP (m/z 225) and [d5]-PhIP (m/z 230) were used, with a 0.10 s dwell time and 0.20 Da span. With a mobile phase flow rate of 150 µL/min, the CH3CN/water gradient was as follows: initial 5% CH3CN, to 15% CH3CN in 10.0 min, then to 60% CH3CN by 15 min, and finally to 100% CH3CN by 17 min. The program then held at 100% CH3CN for 4 min, with a 6 min postequilibration time. The column oven was set to 35 °C. Donors and Milk Samples. Milk samples (10-30 mL) were collected from healthy lactating volunteers, immediately frozen, and then stored at -70 °C. Written consent was obtained from each volunteer, as approved by the University of Guelph Human Subjects Committee. Each volunteer completed a comprehensive questionnaire which recorded lactational and parity data, anthropometric measurements, tobacco smoke, environmental and occupational chemical exposures, and dietary information. Included in the dietary information were foods consumed over the last three meals and information on how the meat, if any, was prepared. Milk Sample Preparation and MCX Extraction. Milk samples were thawed in a 50 °C water bath and sonicated for 5 min. For the method development of the MXC extraction protocol, the five target heterocyclic aromatic amines were added to the bovine milk at four levels for quantification using the method of standard addition with HPLC/UV/fluorescence analysis. For the LC/MS analysis of PhIP in human milk samples, [d5]-PhIP was added to a final concentration of 0.2 ng/mL. Each sample (5.0 mL) was diluted with 0.1 M HCl (5.0 mL). Samples were mixed and then heated again at 50 °C for 5 min. MCX cartridges were preconditioned with methanol (2 mL) followed by water (2 mL). Samples were delivered through the cartridges at a rate of 1 drop/s. The cartridge was then washed with 0.1 M hydrochloric acid (3 mL), then dried with nitrogen before washing with dichloromethane:methanol (2:1) (2 mL) and methanol (2 mL). Again, the cartridge was dried. The final wash was methanol:concentrated ammonia solution:water (20:5:75) (2 mL). The heterocyclic amines were eluted with acetonitrile: concentrated ammonia solution (95:5) (3 mL). The extracts were maintained at 45 °C while dried with nitrogen. Method development samples were resuspended in methanol (50 µL) followed by the addition of 50 µL of 10 mM ammonium acetate buffer, pH 4.0. Aliquots of these solutions (25 µL) were analyzed by the HPLC/UV/fluorescence system. Conversely, samples for analysis by LC/MS were resuspended in methanol (50 µL) and injected (10 µL).
Results Method Development for the MCX Extraction Procedure. The five target heterocyclic aromatic amines chosen for the method development ranged in properties from the highly polar Glu-P-2 to the least polar ARC, with MeIQx, harman, and PhIP having intermediate polarities, based on their chromatographic retention times. Unlike the other heterocyclic aromatic amines in this group, harman does not have an exocyclic amino group and is not mutagenic; however, harman and norharman (another β-carboline) are comutagenic. In the presence of certain nonmutagens, such as aniline and o-toluidine, they react to form potently mutagenic species (27). Table 1 summarizes the recoveries of the five target heterocyclic aromatic amines (at 25 ng/mL) extracted from milk. Recoveries ranged from 85 to 92%, with 3.9-9.4% RSD. The isolation of heterocyclic amines from milk was accomplished on a single mixed-mode cation exchange (MCX) cartridge, using only four wash steps. Analysis of Human Milk Samples. Calibration curves for PhIP were developed using the isotope dilution method. Standards containing PhIP (0.5-100 ng/mL) and [d5]-PhIP (10 ng/mL) were analyzed on a daily basis by
Detection of PhIP in Human Milk
Chem. Res. Toxicol., Vol. 14, No. 11, 2001 1525
Table 1. Recovery of the Five Target Heterocyclic Aromatic Amines from Spiked Bovine Milk with the MCX Extraction Protocola
% recovery (mean) % RSD
Glu-P-2
MeIQx
Harman
PhIP
ARC
85 3.2
86 9.0
90 13
89 5.0
92 9.4
a Bovine milk was spiked with 25 ng/mL of each heterocyclic aromatic amine. Quantification was by the standard addition method, which was performed in triplicate at four addition levels. Analysis was completed by HPLC/UV/fluorescence under the conditions stated in the Experimental Procedures.
Table 2. Level of PhIP Detected in Human Milk Samplesa sample ID
cooked meat in last three meals
PhIP (pg/mL)
103 120 137 140 141 152 159 179 181 202 222d
no yes yes yes yes yes yes yes yes yes no
21 59 ( 12b 13 ( 4b 31 ( 10b 42 15 34 13
