Lebensmittel können mit einer Vielzahl unterschiedlicher chemischer Verbindungen kontaminiert sein. Mittels klassischer chemischer Analytik können gezielt ausgewählte Verbindungen in sehr niedrigen Konzentrationen nachgewiesen werden. Das komplexe Spektrum der möglichen Kontaminanten kann dadurch jedoch kaum vollständig und das Gefährdungspotential nicht immer ausreichend erfasst werden. Daher werden in der Umweltforschung seit etwa 25 Jahren integrierte biologisch-chemische Verfahren entwickelt und eingesetzt, die biologische Effekte von Stoffen der Substanzgemische messen (BRACK 2006). Diese werden als wirkungsorientierte Analyse (effect-directed analysis) oder als Toxizitätsidentifizierung und -auswertung (toxicity identification evaluation) bezeichnet. Neben dem Vorteil der Abschätzung des Summeneffektes einer komplex belasteten Probe beschreibt BRACK (2006) die Vorteile durch Senkung der Analysekosten, da hohe Fixkosten über Geräteanschaffungen (GC-MS / HPLC) reduziert werden können.
In der Untersuchung von Lebensmitteln werden in den letzten Jahren zunehmend wirkungsorientierte Ansätze im Bereich der Screeningverfahren getestet und eingesetzt.
Im Bereich der Dioxinanalytik regelt die europäische Verordnung VO (EG) Nr. 1883/2006 die Analysemethoden für die amtliche Kontrolle. Sie nennt explizit den Einsatz von Screeningverfahren zur Auswahl der Proben mit einem signifikanten Gehalt an Dioxinen.
In den Niederlanden und Belgien sind solche Verfahren seit einiger Zeit etabliert. Als Vorteil gelten der hohe Probendurchsatz und die dadurch bedingte Zeitersparnis.
Ziel dieser Dissertationsarbeit war es, eine neue wirkungsorientierte Analytik von polyzyklischen aromatischen Kohlenwasserstoffen (PAKs) zu entwickeln, an das Untersuchungsgut Fleisch anzupassen und diese Analytik im praktischen Einsatz in der Lebensmittelkontrolle zu erproben.
PAKs stellen eine große, vielfältige Gruppe von chemischen Stoffen dar. Viele Vertreter gelten als mutagen, teratogen und kanzerogen. Ebenso werden Einflüsse auf die Fruchtbarkeit beschrieben (IARC 2005). Sie werden bei unvollständigen Verbrennungen von Fetten gebildet und kontaminieren Lebensmittel über die Umwelt (Ablagerung aus der Luft) und über Bearbeitungsprozesse wie Räuchern, Braten, Grillen und Rösten (GUILLEN
u. SOPELANA 2003). Geschmolzenes Fett, das in die Glut tropft, erhöht die Kontaminationsmenge (WHO 2005). Die Auswahl des Räuchermaterials, Art der Raucherzeugung (extern oder intern), Verfügbarkeit von Sauerstoff und vor allem die eingesetzte Räuchertemperatur haben großen Einfluss auf die Menge von PAKs, die auf und in der Räucherware gebildet wird (SIMKO 2005).
Die wirkungsorientierte Analytik erfolgte mit dem in vitro CALUX Bioassay (chemical activated luciferase gene expression assay), der damit zum ersten Mal in der Analytik auf PAKs eingesetzt wurde. Dieser Bioassay zeigt die Aktivität aller Substanzen einer Probe, die mit dem Aryl Hydrocarbon Rezeptor (AhR) interagieren. Dieser vermittelt die meisten toxischen Effekte von u.a. Dioxinen, polychlorierten Biphenylen und PAKs (GUILLEN u.
SOPELANA 2003). Aufgrund der Spezifität zum AhR ist es nicht möglich, die exakten Substanzen auszumachen, die ursächlich zur Aktivität einer Probe beitragen. Es wird vielmehr ein generelles biologisches (krebserregendes) Potential einer Probe aufgezeigt, das mit der Wirkung der enthaltenen PAKs korreliert. Dies wurde in vitro von MACHALA et al. (2001) und in vivo von SHIMADA et al. (2002) gezeigt.
In kontaminierten Produkten wird stets eine Vielfalt verschiedener PAKs gebildet. Dennoch wurde viele Jahre lang nur die Leitsubstanz Benzo[a]pyren (B[a]P) auf nationaler und europäischer Ebene reglementiert und untersucht. In den letzten Jahren wuchs das Bewusstsein, dass die Bestimmung von B[a]P alleine für eine komplexe Risikobewertung einer Lebensmittelprobe nicht ausreiche, da festgestellt wurde, dass B[a]P nur mit 1 bis 20% zum gesamten kanzerogenen Potential einer Probe beiträgt (SIMKO 2002). Im Jahre 2005 empfahl die Europäische Kommission 15 ausgewählte PAKs, die vom Wissenschaftlichen Ausschuss „Lebensmittel“ als karzinogen eingestuft wurden, weiter zu untersuchen. In der Folge wurde von der europäischen Behörde für Lebensmittelsicherheit (EFSA) aufgrund einer entsprechenden Bewertung durch den Gemeinsamen FAO/WHO-Sachverständigenausschuss für Lebensmittelzusatzstoffe (JECFA) noch ein weiteres PAK der Liste hinzugefügt (WHO 2005). Auf diese, 15+1 priority PAKs genannte Liste1 fokussiert sich zur Zeit die europäische Forschung, obwohl in einem europaweiten
1 Benzo[c]fluorene, Benz[a]anthracene, Cyclopenta[cd]pyrene, Chrysene, 5-Methylchrysene, Benzo[b]fluoranthene, Benzo[k]fluoranthene, Benzo[j]fluoranthene, Benzo[a]pyrene, Indeno[1,2,3-cd]pyrene, Dibenz[a,h]anthracene, Benzo[ghi]perylene, Dibenzo[a,e]pyrene, Dibenzo[a,h]pyrene, Dibenzo[a,i]pyrene, Dibenzo[a,l]pyrene
Ringversuch 2007 festgestellt wurde, dass noch nicht alle Laboratorien diese neuen Anforderungen umsetzen können und nur ein Viertel aller beteiligten Labore in der Lage waren, zufrieden stellend alle 16 PAKs zu bestimmen (SIMON et al. 2008).
Um die Ergebnisse der wirkungsorientierten Analytik von geräucherten Fleischproben mit denen einer herkömmlichen chemischen Analytik vergleichen zu können, wurden in der vorliegenden Arbeit alle Fleischproben parallel zum Bioassay mittels instrumenteller Analytik (GC-MS) untersucht. Dabei wurde auf insgesamt 31 PAKs untersucht, einschließlich der 15+1 EFSA-priority PAKs. Als Untersuchungsgut kamen zum einen Fleischproben (Bauchspeck des Schweins) zum Einsatz, die selbst unter definierten Bedingungen geräuchert wurden. Zum anderen wurden entsprechende Produkte im Groß- und auf dem Wochenmarkt eingekauft und analysiert.
Nach der Adaptation des Bioassays an die Untersuchung auf PAK erfolgte der praktische Einsatz in der Untersuchung der geräucherten Bauchspeckproben. Dabei wurde der Einfluss unterschiedlicher Räucherbedingungen (Räuchertemperatur und Räucherdauer) auf die Entstehung kanzerogener PAKs untersucht und bestimmt, wie sensibel die wirkungsorientierte Analytik mittels Bioassay Änderungen der Räucherbedingungen und damit der Kontaminationsmenge anzeigt.
Kapitel 2
Determination of Polycyclic Aromatic Hydrocarbons in Smoked Pork by Effect-Directed Bioassay with Confirmation by Chemical Analysis
published by Journal of Food Protection, Vol. 71, No. 5, 2008, page 993–999
KERSTIN KUHN,1 BERNHARD NOWAK,1* GÜNTER KLEIN,1 ANDREAS BEHNKE,2 ALBRECHT SEIDEL,2 AND ALFONSO LAMPEN3
1 Institute for Food Quality and Food Safety, University of Veterinary Medicine Hannover, Foundation, Bischofsholer Damm 15, Hannover, D-30173, Germany.
2 Biochemical Institute for Environmental Carcinogens, Prof. Dr. Gernot Grimmer-Foundation, Lurup 4, Großhansdorf, D-22927, Germany.
3 Department of Food Safety, Federal Institute for Risk Assessment, Thielallee 88 – 92, Berlin, D-14195, Germany.
Keywords: smoked pork, PAH, CALUX, bioassay
* Author for correspondence:
Tel: 49-511-856-7319;
Fax: 49-511-856-827319;
E-mail: bernhard.nowak@tiho-hannover.de
Anteil der Erstautorin an der Arbeit:
Versuchsplanung, Probenerstellung und Probenahme, Aufarbeitung und Aufreinigung aller Proben, Durchführung der wirkungsorientierten Analytik, Auswertung der Ergebnisse, Erstellung des Manuskriptes
Copyright䊚, International Association for Food Protection
Determination of Polycyclic Aromatic Hydrocarbons in Smoked Pork by Effect-Directed Bioassay with Confirmation
by Chemical Analysis
KERSTIN KUHN,1BERNHARD NOWAK,1*GU¨ NTER KLEIN,1ANDREAS BEHNKE,2ALBRECHT SEIDEL,2AND
ALFONSO LAMPEN3
1Institute for Food Quality and Food Safety, University of Veterinary Medicine Hannover, Foundation, Bischofsholer Damm 15, 30173 Hannover, Germany;2Biochemical Institute for Environmental Carcinogens, Prof. Dr. Gernot Grimmer-Foundation, Lurup 4, 22927 Großhansdorf, Germany; and3Department of Food Safety, Federal Institute for Risk Assessment, Thielallee 88-92, 14195 Berlin, Germany
MS 07-303: Received 7 June 2007/Accepted 28 November 2007
ABSTRACT
Polycyclic aromatic hydrocarbons (PAHs) are generated during smoke curing and other heating treatments of food and represent a large class of chemical pollutants including a number of carcinogens. At present, PAHs are frequently detected by costly and time-consuming chemical analysis. Effect-directed in vitro cell–based bioassays of contaminants can offer a rapid, sensitive, and relatively inexpensive alternative for screening of contaminants in comparison to instrumental analysis. They enable estimation of total biological activity of all compounds acting through the same mode of binding. The aryl hydrocarbon receptor as a binding site plays an important role in PAH-induced carcinogenesis. The in vitro chemical-activated luciferase expression assay (using conditions to detect PAH) was investigated for its applicability for effect-directed analysis of PAH levels in smoked meat. There was an intra-assay variability of 0 to 15% and a mean coefficient of variation of 25% (3 to 50%) for the cleanup and bioassay analysis of the smoked pork samples. There was a correlation between the total responses of the bioassay and the individual amounts of the PAHs with a high molecular weight. The comparison of 2,3,7,8-tetrachlo-rodibenzo-p-dioxin and benzo[k]fluoranthene used as standard in the in vitro chemical-activated luciferase expression assay resulted in benzo[k]fluoranthene being able to be used as an alternative, nontoxic standard in the bioassay. This bioassay is an applicable effect-directed functional prescreening method for the analysis of PAHs in smoked meat and appears to have potential in being used for food control in the future.
Polycyclic aromatic hydrocarbons (PAHs) are wide-spread environmental contaminants representing a very im-portant group of carcinogens with the representative com-pound benzo[a]pyrene (B[a]P). Teratogenic, mutagenic, and carcinogenic properties have been reported for many PAHs as well as noncarcinogenic effects such as reproduc-tive toxicity(9). PAHs are formed by an incomplete com-bustion of organic material and can be passed on to foods because of industrial and engine combustion emissions con-taminating air, water, and soil (8). Concerning food tech-nology, PAHs are generated during food processing, such as smoke curing, deep frying, and drying. It is likely that PAHs are formed from melted fat that undergoes pyrolysis when dripping onto the heat source (11). The Scientific Committee on Food of the European Commission reports that heat treatment of food is commonly thought to be the major source of contamination by PAHs(5).For nonsmok-ers, food is the main source of exposure to PAHs(5).The contamination of food in the smoking process is highly var-iable and is caused by the type and composition of wood, type of generator (internal or external), oxygen accessibil-ity, temperature of smoke generation, and smoking time(5, 11).The final report on the tasks for scientific cooperation
* Author for correspondence: Tel: 7319; Fax: 49-511-856-827319; E-mail: bernhard.nowak@tiho-hannover.de.
(Task 3.2.12) of the European Union collected data in 2004 on the occurrence of PAHs in food in 13 European coun-tries(2).Thirteen percent of the 1,023 samples in the group,
‘‘traditional smoked products,’’ had an amount above 5 ppb B[a]P, which is the threshold in Regulation 208/2005 of the European Union for smoked meat(2, 3). The mean value in this group was 3.27g/kg, and the median was 0.13g/
kg (3).The Scientific Committee on Food determined that a ‘‘virtually safe dose’’ of B[a]P as marker for the mixture of carcinogenic PAHs in food would be in the range of 0.06 to 0.5 ng B[a]P/kg of body weight per day, which can be converted into a maximum daily intake of 0.04 g/kg B[a]P for an adult (5).The Joint FAO/WHO Expert Com-mittee on Food Additives (JECFA) could not establish a tolerable intake for B[a]P in its 64th report(11).
The aryl hydrocarbon receptor (AhR) mediates the ma-jority of the biological effects of PAHs and halogenated aromatic hydrocarbons–like dioxins. Following binding to the AhR, the ligand-receptor complex is activated and translocated to the nucleus, wherein it binds to a dioxin-responsive element and stimulates transcription of target genes like cytochrome P-450 (CYP) 1A1 (4). Numerous studies support a role for the AhR in mediating the toxicity of PAHs and halogenated aromatic hydrocarbons(5, 8).
In this study, the applicability of the in vitro chemical-activated luciferase expression assay (CALUX) for
detec-TABLE 1. Coefficient of determination (R) of the correlation between the amounts of 27 PAHs and the total CALUX-IEQaof 20 smoked pork belly samplesb
PAH
Sum of benzo[b⫹k⫹j]fluoranthene 0.05–1.15 0.861
Benzo[c]fluorene 0.01–1.05 0.847
Sum chrysene⫹triphenylene 0.07–2.03 0.73
Anthracene 0.46–42.14 0.713
aTCDD induction equivalents; range, 30.4 to 632.1 ng/kg.
bSee Figure 3 for the example B[a]P. Dibenzopyrenes were ana-lyzed in all samples, but were always below the LOQ.
tion and assessment of AhR-mediated activity of PAHs in food probes was evaluated. Such effect-directed analytical methods are widely used in the screening of environmental compounds such as soil, water, and air. With respect to con-sumer safety, the effect-directed analysis is valuable for prescreening food samples because the unknown potency of ingredients and contaminants of meat samples can be determined.
The CALUX cell bioassay uses a rat hepatoma cell line (H4IIE), stably transfected with a luciferase gene contain-ing vector from the firefly, which is under transcriptional control of the dioxin-responsive element(13).Chemical ac-tivation of the AhR results in a production of the enzyme luciferase, generating light on addition of the substrate lu-ciferin. Two types of binding to the AhR are known. Di-oxins and other halogenated aromatic hydrocarbons bind in a relatively stable and persistent manner to the receptor, whereas PAHs are metabolized after some hours, and they can therefore no longer be detected after 24 h of exposure (20).
The potency of a complex mixture such as a meat sam-ple, to cause a defined biological response, is expressed relative to the response of a well-characterized standard or prototypical compound(18).Activities of samples are ex-pressed as bioassay-derived induction equivalents (IEQ) per specified amount of sample (12). In the present study we used 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) and ben-zo[k]fluoranthene (B[k]F) as standards. TCDD was used be-cause of its known affinity to the AhR. B[k]F is known to be a good inducer of CYP enzymes(14),has a high affinity to the AhR, and is described as one of the most potent PAH after incubation of 6 h in the CALUX bioassay(12, 19).
The toxicity is lower in comparison to TCDD, and therefore it could be used as an alternative in the effect-directed anal-ysis.
The aim of this study was to investigate whether the CALUX bioassay could be used as a cheap, easy, and fast effect-directed assay as a prescreening test for the detection of contaminants such as PAHs in smoked food matrices like pork belly. To validate the CALUX bioassay in combina-tion with the extraccombina-tion and cleanup method, a comparison was made between induced equivalents and certain PAHs determined by gas chromatography–mass spectrometry (GC-MS) methodology.
MATERIALS AND METHODS
Chemicals.Cyclohexane (CAS no. 110-82-7), N,N-dimeth-ylformamide (CAS no. 68-122), and acetone (CAS no. 67-64-1) were obtained from Merck (Darmstadt, Germany) in spectroscopy grade. Methanol (CAS no. 67-56-1) was purchased in Pestilyse grade (grade for pesticide analyze) from Roth (Karlsruhe, Ger-many).
TCDD (CAS no. 1746-01-6) was purchased from LGC Promochem (Wesel, Germany). TCDD was dissolved in dimeth-ylsulfoxide (DMSO; CAS no. 67-68-5, spectrophotometric grade, Acros Organics, Geel, Belgium). The final concentration of TCDD was 300 nM. B[k]F (CAS no. 207-08-9, fluorescence grade, Fluka, Steinheim, Germany) was dissolved as 10 mM stock solution in DMSO.
Silica gel 60 (0.06 to 2 mm) was purchased from Roth. All other chemicals used were of high-purity grade.
For the analysis with GC-MS an internal standard mix (d8 -naphthalene, d8-acenaphthylene, d10-acenaphthene, 2F-fluorene, d10-phenanthrene, d10-fluoranthene, d10-pyrene, d12 -benz[a]an-thracene, d12-benzo[b]fluoranthene, d12-benzo[a]pyrene, indeno-[1,2,3-cd]fluoranthene, andd12-benzo[ghi]perylene) was used for the determination of 16 priority PAHs (U.S. Environmental Pro-tection Agency, method 8310), supplemented with further PAHs named in the 64th Report of JECFA and the European Food Safety Authority (see Table 1). The European Food Safety Authority rec-ommends analyzing the 15 E.U. priority PAHs(15),identified as being genotoxic and carcinogenic (benzo[a]anthracene, benzo-[b⫹j⫹k]fluoranthene, B[a]P, benzo[ghi]perylene, chrysene, cyclo-penta[cd]pyrene, dibenz[a,h]anthracene (DBA), dibenzo-[a,e⫹a,h⫹a,i⫹a,l]pyrene, indeno[1,2,3-cd]pyrene, and 5-methyl-chrysene) along with benzo[c]fluorene recommended by JECFA for future monitoring of food.
The dibenzopyrenes were analyzed in all smoked pork belly samples but were always below the limit of quantification (LOQ).
The internal standard mix was provided by the Biochemical Institute for Environmental Carcinogens, Großhansdorf, Germany, which also performed the GC-MS analyses on an Agilent 6890N GC with an Agilent 5973N MSD (Agilent, Santa Clara, Calif.).
The analyses were performed with slight modifications as de-scribed previously (7). They were carried out in the single-ion
monitoring modus, and the quantitation was obtained by compar-ing the area responses of the target ions in relation to the area responses of the internal standards, after a three-point calibration.
Sample extraction and cleanup procedure.Raw pork bel-lies of grade E (premium grade) were purchased and smoked in an FPC100 smoking chamber (Fessmann, Winnenden, Germany) using beech woods shavings. For a broad spectrum of different stages of PAH accumulation, the samples were smoked with dif-ferent standard smoke programs (hot smoke at 65⬚C and cold smoke at 18⬚C, with time differences). Samples were drawn from both the middle and the margins of the pork belly, since PAH accumulation is known to vary between these zones(17).Three samples of pork belly, prepared in a similar way to ours, were bought in a large supermarket in Hannover, Germany, and ana-lyzed in the same way. The (total) 20 samples of pork belly (100 g each) were minced and blended in a chopper (Rondo 1000, Tefal, Offenbach, Germany) and stored frozen (⫺18⬚C), packed in aluminum boxes until the extractions were done.
Four 10-g subsamples—i.e., three for the bioassay and one for the GC-MS—were collected from each of the 20 pork belly samples and prepared for further analysis. Although one subsam-ple would have sufficed for meat control, three were prepared to ensure the reproducibility of the bioassay. Likewise, three addi-tional subsamples were drawn from each of three pork belly sam-ples in order to verify the reproducibility of the GC-MS analysis.
All subsamples were prepared in the same manner, particularly observing equal reaction and storage times. However one differ-ence in the analysis technique did occur: In those samples intend-ed for GC-MS analysis, an internal standard mix including the deuterated isotopes of the PAHs mentioned above was added at the beginning of the extraction.
For both kinds of samples, extraction was performed as rec-ommended by the German Food and Feed Law (§64 LFGB(1)) via the corresponding instructions of the national reference labo-ratory to determine B[a]P in smoked meat products. This method is also described in the 64th Report of JECFA (11) and recom-mended to be used for fatty foods such as meat.
Analysis commenced with an alkaline hydrolysis by boiling the minced sample under reflux conditions for 1 h in 2 M KOH, which had been dissolved in 90% methanol. When the sample was dissolved, a liquid-liquid partition was performed twice, using cyclohexane. Three washing steps with methanol and water (7:3), and one with pure water concluded the first stage of analysis. After the concentration of the solvent a liquid-liquid partition was per-formed twice with 90% N,N-dimethylformamide. The N,N-methylformamide phases were washed with cyclohexane and di-luted with water until the concentration of N,N-dimethylformam-ide was 50% in the solution. After partition with cyclohexane, the solvent was washed with water. The solvent was concentrated to a volume of approximately 3 ml, and a cleanup with a silica gel column was used (300 mm, 10 mm diameter, 5-g conditioned silica gel). The sample was eluted with 110 ml of cyclohexane and concentrated to a final volume of 1 ml. Prior to use, all glass-ware and equipment was rinsed with acetone and cyclohexane.
The cleaned extract was further evaporated under a gentle N2 stream, and 100l of dimethylsulfoxide was added just prior to the complete evaporation of the solvent. Instead of this, samples dedicated for the analysis with GC-MS were shipped in 1 ml of cyclohexane-toluene (9:1) to the Biochemical Institute for Envi-ronmental Carcinogens, where the GC-MS analysis was per-formed.
Cell culture.Rat hepatoma cells (H4IIE), stably transfected with the luciferase reporter gene plasmid pGudLuc1.1(6), were
provided by Prof. Dr. A. Brouwer (BDS, Amsterdam, The Neth-erlands). The cells were grown in DMEM/F-12 with 10% heat-inactivated fetal bovine serum (Mycoplex, PAA, Co¨lbe, Germany) and 1% glutamine. All media and supplements for cell culture were purchased from PAA. Before each test period gentamicin B1 was added for 10 days to the media at 0.5 mg/ml in order to select cells containing the plasmid, which also contained a neomycin resistance marker.
The CALUX bioassay.Cells were seeded in 96-well culture plates (Nunc, Wiesbaden, Germany) at a density of 22,500 per well (200l per well). The margin wells were filled with medium and were not included in the assay. Twenty-four hours after seed-ing, 50l of culture medium containing the test extracts or the standard dilutions was added to the medium in the wells. Expo-sures were done in quadruplicate. Solvent control and TCDD or B[k]F calibration standards (between 0.3 pM and 300 pM, and 1 pM to 1M for B[k]F, respectively) were included in quadrupli-cate on each well plate. The final concentration of DMSO in each well was 0.1% (vol/vol).
The adequate dilution of the extracts in the range of 0.1 to 0.001% was used, which induces a luciferase response in the
The adequate dilution of the extracts in the range of 0.1 to 0.001% was used, which induces a luciferase response in the