Differences in the frequency of micronucleated erythrocytes in humans in relation to consumption of fried carbohydrate-rich food

Differences in the frequency of micronucleated erythrocytes in humans in relation to consumption of fried carbohydrate-rich food, Differences in the frequency of micronucleated erythrocytes in humans in relation to consumption of fried carbohydrate-rich food

Mutation Research 653 (2008) 50–56 Contents lists available at ScienceDirect Mutation Research/Genetic Toxicology and Environmental Mutagenesis journal homepage: www.elsevier.com/locate/gentox Community address: www.elsevier.com/locate/mutres Differences in the frequency of micronucleated erythrocytes in humans in relation to consumption of fried carbohydrate-rich food ă b, Lilianne Abramsson-Zetterberg a,∗ , Anna C Vikstrom b a ¨ ¨ Margareta Tornqvist , Karl-Erik Hellenas a b Livsmedelsverket, National Food Administration, Toxicology Division, Box 622, 751 26 Uppsala, Sweden Department of Environmental Chemistry, Stockholm University, 106 91 Stockholm, Sweden a r t i c l e i n f o Article history: Received December 2007 Received in revised form February 2008 Accepted 12 March 2008 Available online April 2008 Keywords: Micronucleus Transferrin-positive reticulocytes Hemoglobin-adducts Human Acrylamide Carbohydrate-rich food a b s t r a c t The aim of this study was to investigate if consumption of ordinary carbohydrate-rich food prepared in different ways has an impact on chromosome stability, i.e., on the formation of micronucleated young erythrocytes in humans Twenty-four persons, divided into two groups, participated during days in a semi-controlled foodconsumption study One group (low-heated-food-group, LowHF-group) consumed only food boiled in water (max 100 ◦ C) and the other group (high-heated-food-group, HighHF-group) consumed preferentially strongly heated (fried) food From each of the subjects, blood samples were drawn, before and after days The frequency (f) of micronucleated (MN) very young erythrocytes (transferrin-positive reticulocytes, Trf-Ret), fMNTrf-Ret, was determined, and the difference in the frequency, before and after the eating period, was calculated The obtained mean differences for the two groups were compared As an indicator of highly heated food the acrylamide (AA) content in part of the consumed foodstuffs was analysed by use of LC/MS–MS and the AA intake estimated In the blood samples the hemoglobin-adduct levels from AA were analysed as a measure of the internal AA dose The differences between the mean fMNTrf-Ret, before and after the eating period, were −0.15‰ for the LowHF-group and +0.17‰ for the HighHF-group, p < 0.005 (t-test, one-tailed) The mean total AA intake in the HighHF-group during days was estimated to about 3000 ± 450 ␮g per person For the LowHF-group, the mean AA intake was low, 20 ± 10 ␮g per person The lowest dose of AA that caused a significant increase of micronucleated erythrocytes in mice is more than a hundred times higher than the AA level in this study Thus, it is unlikely that the exposure to AA is the major cause behind the observed difference The answer is probably to be found in other compounds produced at the same time during heating of the food © 2008 Elsevier B.V All rights reserved Introduction Probably over an open fire originally, humans have since ancient times used heat to prepare foods They likely sought to make food more tasty, durable, and more easy to masticate Not until the time of Louis Pasteur did it become clear that the improved durability obtained by heating foods was due to the killing of microorganisms Today’s cookbooks are filled with recipes for preparing all sorts of dishes using heat Not the least popular methods for heating food are barbecuing and deep-frying, with French fries and potato chips being examples of common products that are highly heated ∗ Corresponding author Tel.: +46 18 17 57 63; fax: +46 18 17 14 33 E-mail address: liab@slv.se (L Abramsson-Zetterberg) 1383-5718/$ – see front matter © 2008 Elsevier B.V All rights reserved doi:10.1016/j.mrgentox.2008.03.007 Unfortunately, the heating of food has also an adverse effect: polycyclic aromatic hydrocarbons (PAHs), heterocyclic aromatic amines (HAAs), 3-monochloropropanediol (3-MCPD), furan, ethyl carbamate, and acrylamide (AA) are examples of unwanted compounds for which the concentrations increase with increasing cooking temperature PAHs are formed and released during incomplete pyrolysis or combustion of organic material A large part of the PAH contamination of food comes with the flames from the seat of the fire High concentrations of individual PAHs have been found in smoked fish and meat [1] Another type of potent mutagenic compounds is the group of HAAs, among which more than twenty different compounds are identified today Most of the HAAs are produced at very high temperatures, above 200 ◦ C, close to pyrolysis of amino acids and proteins [2] AA, furan, ethyl carbamate, and 3-MPCD have a somewhat different origin than HAA and PAH Furan can be formed in many foods from such common food constituents as ascorbic acid or polyunsaturated fatty acids, at the Boiled fish or meat, veg + boiled potatoes White bread, buns Dinner Extra Boiled fish, veg + boiled potatoes Lunch All lunches were eaten together at the working place Before the study started the participants in the LowHF-group were told not to eat highly fried food, only fresh and boiled (max 100 ◦ C) food The LowHF-group was supplied with fresh potatoes, buns, and white bread Members of the HighHF-group were requested to eat preferentially highly heated food and the participants were supplemented with frozen French fries (1200 g, to be prepared for dinners at home), potatoes crisps (600 g), biscuits (400 g), and crisp bread (600 g) All supplied foodstuffs and all lunches eaten together were analysed with respect to AA content at the Swedish Food Administration The consumption of vegetables (veg.) was almost the same for the two groups Fried fish or meat, veg + French fries Potatoes crisps + biscuits Boiled fish or meat, veg + boiled potatoes White bread, buns HighHF Blood sampling Break-fast Fried fish or meat, veg + French fries Potatoes crisps + biscuits Fried fish, veg + French fries Boiled fish, veg + boiled potatoes Fried sausage and potatoes, veg Fried fish or meat, veg + French fries Potatoes crisps + biscuits Boiled sausage and potatoes, veg Boiled fish or meat, veg + boiled potatoes White bread, buns Fried chicken, veg + French fries Fried fish or meat, veg + French fries Potatoes crisps + biscuits Boiled chicken, veg + boiled potatoes Boiled fish or meat, veg + boiled potatoes White bread, buns Fried fish, veg + French fries Blood sampling Crisp bread + tea/coffee Soft white bread + tea/coffee Crisp bread + tea/coffee Soft white bread + tea/coffee Crisp bread + tea/coffee Soft white bread + tea/coffee Blood sampling Day LowHF HighHF Day LowHF HighHF Day LowHF HighHF Day LowHF HighHF Day LowHF Group Table The menu during days of the study Blood sampling L Abramsson-Zetterberg et al / Mutation Research 653 (2008) 50–56 51 normal food canning/sterilisation temperature [3] Due to the high volatility of furan, heat-processed canned and jarred foods accumulate the highest levels of this chemical [4] Heat-induced formation of ethyl carbamate and 3-MCPD may in both cases include precursors formed by yeast fermentation (e.g., ethanol and glycerol, respectively), and these compounds are consequently present in bakery products [5,6] Concerning AA, high concentrations have been found in carbohydrate-rich food such as potatoes and cereal products that contain both free asparagine and reducing sugar Already at temperatures far below 200 ◦ C, AA is produced, with an optimum of about 160–180 ◦ C, i.e., the common temperatures for baking and frying [7] The occurrence of AA in food was first published by Tornqvist and colleagues [8] They showed an AA content in fodder aimed for laboratory animals Two years later they published data about high AA levels, up to a mg per kg in, e.g., fried potatoes [9] Followup studies verified the findings [10], and data bases on AA levels in many different foodstuffs are available today [11,12] As early as 1994 the genotoxicity of AA was verified by IARC (International Agency for Research on Cancer, WHO organ) and the compound was classified as a probable human carcinogen, Group 2A [13] AA is probably not the only genotoxic chemical that is produced when heating carbohydrate-rich foodstuffs There may be other chemicals that have not yet been studied and that may be of equal concern The EU-funded Heatox project (Heat-induced food toxicants, identification, characterisation and risk minimisation at http://www.heatox.org) has recently listed around 800 compounds that are either products of the Maillard reaction [14,15] or of lipid oxidation [16] A positive correlation in cancer epidemiology is seen as an ultimate proof of carcinogenicity for a given factor in humans Several epidemiological studies published during the period 2003–2007 show no positive association between dietary intake of AA or selected AA-containing foods and increased risks for cancer [17–21] However, a recently published study did find a positive association between dietary AA intake and an increased risk for certain types of cancer (uterus and ovary) in postmenopausal women [22] Whilst this study provides epidemiological indications that dietary AA may increase the cancer risk in humans, it does not establish an unequivocal causal relationship as such However, the sensitivity of epidemiological studies is sometimes too low to detect the relatively moderate cancer risk increment from AA (as expected on the basis of toxicological studies) [23] In epidemiological studies the AA intake is often estimated from food-frequency questionnaires The reported intake of AA can be verified by measuring hemoglobin-adducts of AA (Hb-AA), which is a biomarker for the internal dose over the last months Several studies have been published that show a weak correlation between the Hb-AA level and the reported food intakes [24–28] Biomarkers of effects, e.g., micronucleated cells in humans are a helpful tool when the sensitivity of epidemiological studies is not sufficiently high to establish a risk [29] Some years ago, Abramsson-Zetterberg et al [30] demonstrated that by separating the very youngest erythrocytes (transferrin-positive reticulocytes (Trf-Ret), in peripheral blood), it is possible to detect micronucleated reticulocytes in humans The occurrence of transferrin receptors on the surface of the very youngest erythrocytes and the use of antibodies make this separation possible Furthermore, a flow cytometer instead of a microscope allows analysis of many cells in the same experiment, which is a prerequisite for a high sensitivity Half a million cells is a common number of analysed Trf-Ret from each sample In the present study we have analysed and compared the frequency of micronucleated Trf-Ret (fMNTrf-Ret) between two groups of persons in a semi-controlled food-consumption study, 52 L Abramsson-Zetterberg et al / Mutation Research 653 (2008) 50–56 Table This table shows the estimated total intake of acrylamide (AA) during the eating period of days, the level of hemoglobin AA-adducts (Hb-AA, in pmol/g globin), the number of transferrin-positive reticulocytes (Trf-Ret) analysed, the number of micronuclei (MN) registered, and the MN frequency (fMN) (‰) both before and after the eating period Food treatment Id Intake of AA (␮g) Before eating period Low-heated-food 2◦ 13◦ 15 18 19 21 23 24 30 15 25 20 35 20 15 15 15 15 10 30 305 43 42 36 63 328 71 103 56 58 64 55 20 102 3◦ 6◦ 10 11 12 14◦ 16 17 20 22 3500 3300 2000 2700 3100 3400 3200 3500 3000 3300 3300 2500 58 103 183 56 41 61 57 122 64 64 43 71 3050 77 Hb-AA Mean High-heated-food Mean Trf-Ret After eating period MN fMN Hb-AA 523,758 574,184 398,008 614,027 520,347 576,103 607,629 615,201 609,560 589,128 676,371 544,390 1277 893 1011 524 947 1336 376 417 1063 765 820 858 2.43 1.55 2.53 0.85 1.82 2.31 0.62 0.68 1.74 1.30 1.21 1.57 297 41 41 33 64 283 77 99 58 60 61 48 669,395 592,808 557,580 619,085 590,552 612,309 582,642 625,139 548,764 413,679 585,538 611,969 644 168 770 597 885 1002 1319 785 854 507 520 1005 1.55 97 0.96 0.28 1.38 0.96 1.50 1.63 2.26 1.25 1.55 1.22 0.89 1.64 106 135 200 99 93 135 128 168 103 109 126 125 1.29 127 ** Trf-Ret Difference MN fMN Hb-AA fMN 590,921 578,947 589,483 590,018 526,732 577,911 598,167 623,251 583,068 336,362 612,535 588,204 1615 941 1603 297 702 888 335 433 913 313 842 751 2.73 1.62 2.71 0.50 1.33 1.53 0.56 0.69 1.56 0.93 1.37 1.28 −8 −2 −1 −3 −45 −4 2 −3 −7 0.29 0.07 0.18 −0.35 −0.49 −0.78 −0.06 0.02 −0.18 −0.37 0.16 −0.30 1.40 −5 −0.15 575,312 597,328 603,519 613,095 592,478 600,192 582,461 593,342 534,696 626,181 589,724 600,866 735 295 849 553 908 1007 1434 1147 795 1092 607 996 1.28 0.49 1.40 0.90 1.53 1.67 2.46 1.93 1.48 1.74 1.03 1.65 48 32 17 43 52 74 71 46 39 45 83 54 0.31 0.21 0.03 −0.06 0.03 0.04 0.20 0.68 −0.07 0.52 0.14 0.02 1.46 50*** +0.17** *** Also listed are the differences in Hb-AA and fMN before and after Id = identification of study participants p < 0.005 and p < 0.0005 (Student’s t-test, one-tailed) When the smokers (indicated by ◦ ) are excluded from the calculation, the mean fMN difference between the two groups is still significant, p < 0.01 (Student’s t-test, one-tailed) one group eating medium-heated (max 100 ◦ C) food and one eating preferentially highly heated food In addition, Hb-AA levels were measured to verify the difference in AA intake between the groups Material and methods 2.1 Chemicals 2.1.1 Analysis of acrylamide in food Acrylamide (CAS 79-06-1) (assay (GC) ≥ 99.9%) was obtained from Merck, Darmstadt, Germany; deuterium-labelled acrylamide-d3 was obtained from Polymer Source Inc Dorval, Quebec, Canada; the columns used for solid-phase extraction (SPE-columns) were Isolute Multimode, and ENV+, from IST, Hengoed, Mid Glamorgan, UK, the analytical LC column was Hypercarb (5 ␮m, 50 mm × 2.1 mm) from Thermo Electron Corp., UK 2.1.2 Analysis of PAH in food Internal standard was per-deuterated PAH-II from Nist, Standard Reference Material 2270 containing Pyrene-d10, Perylene-d12, Benzo(ghi)perylene-d12 (LGC ă Promochem), Chrysene-d12 from Dr Ehrenstorfer-Schafers (LGC Promochem) PAH ă standards were PAH–mix 68 from Dr Ehrenstorfer-Schafers (LGC Promochem) Benzo(j)fluoranthene, 5-Methylchrysene, Dibenzo(a,l)pyrene, Dibenzo(a,i)pyrene, Dibenzo(a,l)pyrene, Dibenzo(a,h)pyrene, Benzo(c) uorene, from Dr Ehrenstorferă Schafers (LGC Promochem) Injection standard: bb-binaphthyl (external standard) in cyclohexane (0.04 ␮g/ml) 2.1.3 Micronucleus test Magnetic beads (CELLectionTM Pan Mouse IgG Kit) were purchased from Dynal, Oslo, Norway The mouse anti-human CD71 antibody was purchased from DakoCytomation, Glostrup, Denmark The fluorescent dye Hoechst 33342 was purchased from Sigma, Aldrich, Sweden, and Thiazole Orange from Molecular probes, Eugene, Oregon, USA 2.1.4 Analysis of hemoglobin-adducts from AA Pentafluorophenyl isothiocyanate (>95%, Fluka, Buchs, Switzerland) was purified on a Sep-Pak silica cartridge [31] Myoglobin, from horse skeletal muscle (Sigma–Aldrich Chemie GmbH, Schnelldorf, Germany) was precipitated in acidic acetone solution Formamide (Scharlau Chemie S.A., Barcelona, Spain) was purified by extraction with n-pentane All other chemicals used were of analytical grade 2.2 Design of the study Twenty-four healthy persons, 24–60 (mean = 46) years old, participated in this study Five were smokers The subjects were randomly divided into two groups, women and men in each of the two groups Each of the subjects donated blood twice, before and after an “eating period” of days One of the groups (low-heatedfood-group, LowHF-group) ate during days medium-heated (max 100 ◦ C) and fresh vegetables and the other group (high-heated-food-group, HighHF-group) eating preferentially strongly heated (fried) and fresh vegetables The menus are shown in Table The frequency of micronuclei and the Hb-AA levels were analysed in all the blood samples As an indicator of heated food the AA levels in part of the foodstuffs consumed were analysed The subjects belonging to the LowHF-group were instructed to minimize the consumption of high-heated food during the study and the other group, the HighHF-group, should choose preferentially fried food Since all participants worked at the same place, it was possible to eat the lunches together and see that the instructions were followed Concerning the food eaten at home, before and after work, the subjects belonging to the HighHF-group were supplied with a certain amount of frozen French fries, potato crisps, biscuits, and crisp bread The other group, the LowHF-group, was supplied with fresh potatoes, buns, and white bread One package of each of the supplied foodstuff (for both groups) was kept at the working place for analysis of the AA content All participants kept a food diary over days The reported consumption of vegetables and fruits were about the same in both groups Concerning coffee, almost all participants reported daily consumption Since the participants in the LowHF-group were instructed to keep the coffee intake low, the consumption of coffee in this group was less than in the HighHF-group To estimate the AA intake during these days, all lunches were pooled each day (i.e., for each day meals, one aimed for the LowHF-group and one for the HighHFgroup) and the homogenated pools were separately analysed by LC/MS–MS at the National Food Administration (Table 3) The AA content in all other foodstuffs supplied to the two groups, to be eaten at the working place or at home, were analysed separately Some of the foods (two dishes of potato crisps, two dishes of French fries, one dish of boiled and one of fried sausage) were also analysed for PAH The study has been approved by the Ethics Committee, Faculty of Medicine, Uppsala University (Dnr Ups 02-094) 2.3 The micronucleus assay in human transferrin-positive reticulocytes, Trf-Ret Part of the blood samples was prepared for analysis of micronucleated erythrocytes, the other part for measurement of Hb adducts Concerning the micronucleus L Abramsson-Zetterberg et al / Mutation Research 653 (2008) 50–56 test, the cells were prepared in accordance with earlier described methods [30,32] In brief, from each participant a mean of 1.5 ml whole blood was washed and mixed with 45 ␮l of a suspension of magnetic beads (CELLection TM Pan Mouse IgG Kit) pre-coated with mouse anti-human CD71 antibody Since it is only the very youngest erythrocytes and part of the nucleated cells that express CD71 antigen on the cell surface, these cells are captured with the antibody After the enrichment procedure of the Trf-Ret, these cells were removed from the beads by DNAse The collected cells (Trf-Ret) were sphered with 400 ␮l detergent solution (SDS) during one and fixed with about ml freshly prepared 2% formalin in PBS at room temperature The cells were stored in the fixation solution at room temperature for months Finally, the day before the flow-cytometer analysis, all cell samples were centrifuged and the pellet was re-suspended in 0.5 ml fresh staining solution (60 ml PBS with 14 ␮l Hoechst 33342, HO342, (500 mM) and 12 ␮l Thiazole Orange, TO, (1 mg/ml) The tubes were incubated at 37 ◦ C for 60 The stained samples were left overnight at +4 ◦ C The coded and stained samples were analysed at a rate of 500–1000 cells/s on a FACSVantage SE flow cytometer (BD Immunocytometry systems, Sunnyvale, CA) equipped with an argon-ion laser (Enterprise II, Coherent, Santa Clara, CA) operating at both multiline 350 (UV) and 488 nm For each analysed cell in the flow cytometer, information about the size, structure, and amount of emitted light (from the DNAand RNA-binding dyes) is transformed to electric signals Signals for forward scatter (FSC, linear), side scatter (SSC, log), TO fluorescence (log), and HO342 fluorescence (log) were collected A mean of 550,000 Trf-Ret were analysed per sample CellQuest software (BD) was used for data acquisition 2.4 Analysis of hemoglobin-adducts The analysis of Hb adducts from AA (Hb-AA) was performed according to the N-alkyl-Edman method [33] Briefly, globin was first precipitated from erythrocytes Then samples of 50 mg globin were dissolved in formamide (1.5 ml) and derivatised with pentafluorophenyl isothiocyanate (8 ␮l) to detach the N-terminal valines, to which AA binds The detached derivative, the pentafluorophenylthiohydantoin (PFPTH) of N-(2-carbamoylethyl)-valine was isolated and purified by extraction The quantification of the level of the Hb adducts from AA was performed by gas chromatography/tandem mass spectrometry (GC–MS/MS) in the negative ion/chemical ionisation mode, using corresponding deuterium-substituted PFPTH analyte, synthesised from (2 H7 )valine, as internal standard Reference globin with known adduct level from AA to N-terminal valine was used for calibration [34] The calibration curve (R2 = 0.992) included samples in the range of 0–400 pmol/g globin The limit of quantification (LOQ) was estimated to be pmol/g globin 2.5 Analysis of the acrylamide content in the food Analysis of AA was performed by liquid chromatography tandem massspectrometry (LC–MS/MS) using electrospray ionisation Briefly, homogenised samples were extracted with water at room temperature, and deuterium-labelled AA was added as an internal standard Two types of solid-phase extraction columns were used to achieve an extract pure and concentrated enough to enable reliable quantification of AA in solid matrices down to ␮g/kg [35] The method was subjected to a multi-laboratory collaborative trial validation study including bakery ware and potato products with AA contents in the range 20–9000 ␮g/kg [36] It was concluded that the method was unbiased, and the method precision (CV of 3–9% and 5–13% for repeatability and reproducibility, respectively) was better than prescribed by international standards 2.6 Analysis of the PAH content in the food Different PAHs were separated by gas chromatography using a capillary column and quantified through mass spectrometric detection (GC–MS) Samples were saponified with potassium hydroxide, extracted with cyclohexane After washing the extract with methanol/water it was purified on two SPE silica columns prior to GC–MS analysis The method is based on published procedures [1,37] with some modifications The applied method was validated for meat, fish, and vegetable oil in the measurement range 0.3–40 ␮g/kg The validation included the 16 different PAHs listed by the EU Commission [38] and some other compounds 2.7 Statistical methods The differences in fMNTrf-Ret (fMN-diff) before and after the eating period were calculated for each of the participants The obtained mean fMN-diff for the two groups (HighHF- and LowHF-group) were then compared with a Student’s t-test A comparison was also made (fMN-diff) with the Student’s t-test between the two groups with no smokers included The same Student’s t-test procedure was used when the differences in Hb-adducts were calculated 53 Fig The difference, in each participant, between the frequencies of micronucleated transferrin-positive reticulocytes (fMN) before and after an eating period of days The participants were divided over two groups, LowHF- (low-heated-foodgroup) and HighHF-group (high-heated-food-group) The LowHF-group ate during days only boiled (max 100 ◦ C) carbohydrate-rich food and the other group (HighHFgroup) ate preferentially strongly heated (fried) carbohydrate-rich food The fMN differences in the two groups are significantly different from each other, p < 0.005 (Student’s t-test, one-tailed) When the five smokers (◦ = smokers) were excluded from the calculation the mean fMN differences were −0.13‰ in the LowHF-group and +0.10‰ in the HighHF-group which also is significant, p < 0.01 (Student’s t-test, one-tailed) Results The mean differences between the frequencies of micronucleated Trf-Ret, fMNTrf-Ret, for each person, before and after the food-consumption trial, were −0.15‰ for the LowHF-group and +0.17‰ for the HighHF-group (see Table and Fig 1) These values are significantly different from each other, p < 0.005 (t-test, onetailed) When the five smokers were excluded from the calculation the differences were −0.13‰ and +0.10‰, respectively, which is also significant, p < 0.01 (t-test, one-tailed) The mean fMNTrf-Ret for both groups together (LowHF and HighHF) were 1.42‰ before and 1.43‰ after the eating period These frequencies are in the normal range [32] The participants were divided by chance into two groups and the fMNTrf-Ret before the eating period started were 1.29‰ and 1.55‰ for the HighHF- and LowHF-group, respectively (Table 2) Excluding the smokers from the calculation, the fMNTrf-Ret were 1.40‰ and 1.39, respectively (calculated from Table 2) The AA contents in the different foodstuffs are given in Table French fries included in the lunch meals eaten at the working place contained AA in the range of 700–1700 ␮g/kg, with a mean level of 1100 ␮g/kg This mean value was also used when the AA intake from the French fries consumed at home was calculated The AA levels in fried lunch dishes, French fries excluded, were low (21–25 ␮g/kg), with the exception of one dish including fried potato (190 ␮g/kg) AA levels in all foods supplied to the LowHF-group were below the quantification limit of the analytical method (i.e.,