Chemical Hazards from Poultry Meat: Cooking, Digestion

Chemical hazards from poultry meat: cooking, digestion,

absorption and toxicity

Thesis submitted to University do Porto, in partial fulfilment of requirements for the PhD degree in Sustainable Chemistry

Maria Madalena Costa Sobral

Under the supervision of: Associate Professor Isabel Maria Pinto Leite Viegas Oliveira Ferreira

And the co-supervision of: Doctor Miguel Ângelo Rodrigues Pinto de Faria Doctor Sara Cristina da Silva Cunha

Porto May, 2020

©AUTHORIZED THE COMPREHENSIVE REPRODUCTION OF THIS THESIS ONLY FOR RESEARCH PURPOSES THROUGH A WRITTEN DECLARATION OF THE INTERESTED PART THAT IS COMMITTED TO SUCH PLEDGES

This work was supported by Fundação para a Ciência e Tecnologia (FCT) through PhD grant (PD/DB/114581/2016), by FEDER - Fundo Europeu de Desenvolvimento Regional funds through the COMPETE 2020 - Operational Programme for Competitiveness and Internationalisation (POCI), and by the DIETImpact project (PTDC/SAU-NUT/30322/2017 and POCI-01-0145-FEDER-030322). It was also supported by UID/QUI/50006/2013 and UID/QUI/50006/2019.

iii

The experimental work presented in this thesis was undertaken under the PhD Programme in Sustainable Chemistry (REQUIMTE), hosted by University of Porto, Nova Lisbon University, and University of Aveiro. The experimental work was mainly performed in the Laboratory of Bromatology and Hydrology, Department of Chemical Sciences, Faculty of Pharmacy, University of Porto (Portugal). iv

AGRADECIMENTOS

No culminar desta etapa gostaria de agradecer a todas as pessoas que de algum modo contribuíram durante o meu percurso académico ao longo destes anos. Em primeiro lugar quero expressar toda a minha gratidão à Professora Isabel por ter aceitado ser minha orientadora, tornando possível este doutoramento. Agradeço toda a ajuda, orientação, dedicação, pronta disponibilidade em todas as etapas, e acima de tudo por ter acreditado em mim. Quero agradecer também ao Doutor Miguel e à Doutora Sara, pela co-orientação, por todo o apoio ao longo destes anos, pelas sugestões e discussões científicas no decorrer dos trabalhos, mas acima de tudo pelos conhecimentos transmitidos. À Professora Susana por todo o apoio, discussões de trabalho, e transmissão de conhecimento que ajudaram a enriquecer esta tese. Às minhas meninas, muito obrigada por me terem acolhido no vosso grupo, por todas as conversas científicas e pessoais, pela boa disposição, e por todos os nossos almoços e “catraiadas” para celebrar os nossos sucessos! De uma forma geral, agradeço a todos os colegas de laboratório de Bromatologia e Hidrologia pelo auxílio em todas as atividades desenvolvidas. Me gustaría agradecer a Toñi y Roberto por haberme recibido en vuestro laboratorio en la Universidad de Almería. También a los demás miembros del grupo de Química Analítica de Contaminantes. Muchas gracias por la integración en el grupo, por la simpatía, por toda la ayuda, disponibilidad y conocimiento compartido. Los meses que estuve ahí me dejaron buenas memorias y enseñanzas muy importantes para mí, tanto profesionalmente como personalmente. ¡Muchas gracias a todos! Não há palavras suficientes para agradecer aos pilares da minha vida, os meus pais e irmão, pelo apoio incondicional, paciência, compreensão e por se orgulharem sempre de mim. Ao meu companheiro, pelo ser fantástico e inigualável que é, pelo incansável apoio, carinho, amizade, compreensão e sobretudo pela calma e serenidade que sempre me transmite nos momentos menos bons. Muito obrigada por seres o meu porto seguro e por me ajudares cada dia a ser uma pessoa melhor.

Um sentido obrigada a todos!

ABSTRACT

Poultry meat is rich in valuable nutrients, such as high-quality proteins, vitamins, minerals, and polyunsaturated fatty acids. Chicken is one of the most consumed poultry meat worldwide with its production and supply raising over the years, which has also increased the occurrence of chemical hazards as cocciodiostats and antibiotic residues, used to fight infections in animals, as well as mycotoxins that can enter the food chain via fungal infection of crops and consequent carry-over to humans. The cooking of meat can also contribute to the formation of deleterious compounds due to oxidation of nutrients. Some culinary practices, such as herbs addition, marinades, or roasting bags have been tested to avoid formation of hazardous substances during cooking while keeping the nutritional value of meat. Nonetheless, there is no information concerning their impact on mycotoxins or antibiotics, nor on lipid and protein oxidation of poultry meat. Thus, this thesis aimed to evaluate: i) the influence of adding ingredients (oregano/beer) on lipid and protein oxidation after cooking (oven/microwave) and in vitro digestion of chicken meat. Therefore, five oxidation markers - malondialdehyde (MDA), 4-hydroxy-2- nonenal (HNE), hexanal (HEX), carbonyls, and Schiff bases – as well as free amino acids and total fatty acids content were measured; ii) moreover, the stability of 14 antibacterial and coccidiostats drugs (ACDs) during cooking and in vitro digestion was also ascertained concerning the addition of oregano and/or beer, and, iii) the behaviour of 10 prevalent mycotoxins was also evaluated during cooking and digestion of chicken, with/without roasting bags and/or commercial mixture of herbs. As the amount of compounds present in cooked meat does not directly reflect the actual amount available for absorption and toxicity, in vitro digestion studies, using the INFOGEST protocol, were also performed for all chemical hazards. Moreover, the natural co-occurrence of mycotoxins and the lack of information on their absorptive and toxicological behaviour when in combination has instigated the evaluation of the transport of four prevalent mycotoxins - aflatoxin B1 (AFB1), deoxynivalenol (DON), fumonisin B1 (FB1), and ochratoxin A (OTA), - isolated and in mixture, across gastric NCI- N87 and intestinal Caco-2 monolayers, as well as their binary toxicological interactions to determine the combinations as additive, antagonistic or synergistic concerning toxicity using as models the Caco-2 (intestinal) and HepG2 (hepatic) cells. As to results, oregano prevented MDA, HEX, and HNE formation, while beer seemed not to influence their formation. The ACDs were stable during cooking, except amoxicillin, chlortetracycline and tylosin (reductions >50%), being molecular rearrangement and dechlorination reactions the most probable transformations derived from cooking. Adding vii

oregano did not benefit their reductions. Mycotoxins were reduced up to 60% by cooking in the presence of herbs; the roasting bags did not confer any advantage. After in vitro digestion, MDA, carbonyls, and Schiff bases increased, regardless of the culinary practice, while HNE and HEX values were reduced. ACDs exhibited maximum bioaccessibility of 60% for all compounds probably due to drug-bile salts interaction suggesting that the bioaccessibility prediction must not be based only on ACDs determination in the free form. In the case of mycotoxins, they exhibited the highest bioaccessibility in digested samples cooked with herbs. However, despite their higher bioaccessibility, as cooking with herbs highly reduced their contents, the exposure to these mycotoxins after digestion is lower in samples with added herbs. Besides the increase of oxidation markers, their lowest contents were observed after cooking with oregano, thus, it should be recommended as a mitigation strategy to avoid the formation of oxidation products during cooking, as well as diminishing their formation during in vitro digestion. Herbs also showed to be an effective strategy to reduce mycotoxins in meat, while in the case of ACDs oregano’s influence was not clear. Moreover, more strategies to reduce ACDs and mycotoxins should be the subject of future research together with more accurate methodologies to assess their bioaccessibility. Concerning gastrointestinal transport assays of mycotoxins, different absorptive patterns on both epithelia were found between molecules isolated or in mixture highlighting that their co-ocurrence and consequent co-exposure considerably impacts human absorption and toxicity. When evaluating the cytotoxicity of combined mycotoxins, synergism or antagonism effects on both Caco-2 and HepG2 cells were observed. Synergistic strengths as high as a dose reduction index of 10 for AFB1-DON were observed in hepatic cells. Taken together our findings indicate that the toxicological effects differ regarding the type of mycotoxins used for combinations and the stronger synergistic effect was observed for DON containing mixtures in both cells. Therefore, even though DON has not been classified as to its carcinogenicity to humans, this mycotoxin may present a serious threat to health, mainly when co-occurring in the environment.

Keywords: lipid oxidation products, protein oxidation, antibiotics, mycotoxins, in vitro digestion, in vitro absorption, combined cytotoxicity

viii

RESUMO

A carne de aves é nutricionalmente importante, sendo rica em proteínas, vitaminas, minerais e ácidos gordos poliinsaturados. O frango é uma das carnes de aves mais consumidas em todo o mundo, no entanto a sua elevada produção e fornecimento ao longo dos anos aumentou a ocorrência de riscos químicos como resíduos de antibióticos, utilizados no combate a infecções em animais, e micotoxinas que podem entrar na cadeia alimentar através de infecção fúngica das culturas e consequente transferência para o homem. O processo de cocção da carne também pode contribuir para a formação de compostos deletérios devido à oxidação de nutrientes. Algumas práticas culinárias, como a adição de ervas, marinadas ou o uso de sacos de assar, tem sido testadas para evitar a formação de oxidação de substâncias perigosas durante a cocção, mantendo o valor nutricional da carne. No entanto, não há informação disponível sobre o seu impacto nas micotoxinas ou antibióticos, nem na oxidação de lípidos e proteínas da carne de aves. Assim, esta tese teve como objetivo: i) avaliar a influência da adição de ingredientes (orégãos/cerveja) na oxidação de lípidos e proteínas após cocção (forno/microondas) e digestão in vitro de carne de frango. Para tal, foram medidos cinco marcadores de oxidação - malondialdeído (MDA), 4-hidroxi-2-nonenal (HNE), hexanal (HEX), carbonilos e bases de Schiff - assim como aminoácidos livres e conteúdo total de ácidos gordos; ii) além disso, também foi verificada a estabilidade de 14 antibióticos e coccidiostáticos (ACD) ao longo da cocção e digestão relativamente à adição de orégãos e/ou cerveja: iii) e avaliar o comportamento de 10 micotoxinas prevalentes também foi avaliado durante a cocção e digestão de frango, com/sem sacos de assar e/ou adição de uma mistura comercial de ervas. Como a quantidade de compostos presente na carne cozinhada não reflete diretamente a quantidade real disponível para absorção e toxicidade, estudos de digestão in vitro, utilizando o protocolo INFOGEST, também foram realizados para todos os riscos químicos. Além disso, a coocorrência natural de micotoxinas e a falta de informação sobre a absorção e toxicidade quando combinadas, instigaram a avaliação do transporte de quatro micotoxinas prevalentes - aflatoxina B1 (AFB1), desoxinivalenol (DON), fumonisina B1 (FB1) e ocratoxina A (OTA), - isoladas e em mistura, usando monocamadas gástricas NCI- N87 e intestinais Caco-2, bem como as suas interações toxicológicas em combinações binárias para determinar interações aditivas, antagónicas ou sinérgicas, usando como modelos as células Caco-2 (intestinal) e HepG2 (hepática). Quanto aos resultados, os orégãos impediram a formação de MDA, HEX e HNE, enquanto a cerveja pareceu não influenciar a sua formação. Os ACDs foram estáveis

durante a cocção, exceto a amoxicilina, a clortetraciclina e a tilosina (reduções >50%), cujas reduções podem dever-se ao rearranjo molecular e reações de descloração. A adição de orégãos não beneficiou as suas reduções. As micotoxinas foram reduzidas ao cozinhar até 60% na presença de ervas, já os sacos de assar não conferiam nenhuma vantagem. Após a digestão in vitro, os níveis de MDA, carbonilos e bases de Schiff aumentaram, independentemente da prática culinária, enquanto os valores de HNE e HEX foram reduzidos. As ACDs exibiram bioacessibilidade máxima de 60% para todos os compostos provavelmente devido à interação com sais biliares, sugerindo que a previsão de bioacessibilidade não deve ser baseada apenas na determinação de ACDs na forma livre. No caso das micotoxinas, as amostras digeridas cozidas com ervas exibiram maior bioacessibilidade. No entanto, apesar de estarem mais bioaccessíveis, como a cocção com ervas foi eficaz na redução das micotoxinas a exposição a estas toxinas será inferior nas amostras com adição de ervas. Apesar do aumento dos marcadores de oxidação, os menores índices foram observados após a cocção com orégãos, pelo que a sua adição deve ser recomendada como estratégia de mitigação para evitar a formação de produtos de oxidação durante a cocção, bem como diminuir sua formação durante a digestão in vitro. As ervas também mostraram ser uma estratégia eficaz para reduzir as micotoxinas na carne, enquanto no caso dos ACDs a influência dos orégãos não foi clara. Além disso, mais estratégias para reduzir as ACDs e as micotoxinas devem ser objeto de futura investigação, juntamente com metodologias mais precisas para avaliar a bioacessibilidade. Em relação aos ensaios de transporte e citotoxicidade de micotoxinas, foram encontrados diferentes padrões de absorção nos dois epitélios durante o transporte isolado ou em mistura evidenciando que a coocorrência e consequente coexposição destes compostos tem impacto considerável na absorção e toxicidade humana. Ao avaliar a citotoxicidade das micotoxinas combinadas, efeitos de sinergismo ou antagonismo foram estudados nas células Caco-2 e HepG2. Forças sinérgicas tão altas quanto um índice de redução de dose de 10 para AFB1-DON em células hepáticas foram observadas. Em suma, os nossos resultados indicam que os efeitos toxicológicos diferem em relação ao tipo de micotoxinas usadas para combinações e o efeito sinérgico mais forte foi observado nas misturas contendo DON em ambas as células. Portanto, embora o DON não tenha sido classificado quanto à sua carcinogenicidade para seres humanos, essa micotoxina pode representar uma séria ameaça à saúde, principalmente quando coocorre no ambiente.

Palavras-chave: produtos de oxidação lipídica, antibióticos, micotoxinas, digestão in vitro, absorção in vitro, toxicidade combinada

LIST OF PUBLICATIONS

PUBLICATIONS IN INTERNATIONAL PEER-REVIEWED JOURNALS

Sobral MMC, Casal S, Faria MA, Cunha SC, Ferreira IMPLVO. Influence of culinary practices on protein and lipid oxidation of chicken meat burgers during cooking and in vitro gastrointestinal digestion. Food and Chemical Toxicology. 2020. 141:111401. (Impact factor 2018: 3.78)

Sobral MMC, Romero-Gonzalez R, Faria MA, Cunha SC, Ferreira IMPLVO, Garrido- Frenich A. Stability of antibacterial and cocciodiostat drugs on chicken meat burgers upon cooking and in vitro digestion. Food Chemistry. 2020. 316:126367. (Impact factor 2018: 5.39)

Sobral MMC, Faria MA, Cunha SC, Miladinovic B, Ferreira IMPLVO. Transport of mycotoxins across human gastric NCI-N87 and intestinal Caco-2 cell models. Food and Chemical Toxicology. 2019. 131:110595. (Impact factor 2018: 3.78)

Sobral MMC, Cunha SC, Faria MA, Martins ZE, Ferreira IMPLVO. Influence of oven and microwave cooking with the addition of herbs on the exposure to multi-mycotoxins from chicken breast muscle. Food Chemistry. 2019. 276:274-84. (Impact factor 2018: 5.39)

Sobral MMC, Faria MA, Cunha SC, Ferreira IMPLVO. Toxicological interactions between mycotoxins from ubiquitous fungi: impact on hepatic and intestinal human epithelial cells. Chemosphere. 2018. 202(C):538-48. (Impact factor 2018: 5.11)

Sobral MMC, Cunha SC, Faria MA, Ferreira IMPLVO. Domestic cooking of muscle foods: impact on composition of nutrients and contaminants. Comprehensive Reviews in Food Science and Food Safety. 2018. 17(2):309-33. (Impact factor 2018: 8.74)

BOOKS

Sobral MMC, Cunha SC, Faria MA, Ferreira IMPLVO. Aflatoxins in food and feed: occurrence, legislation and mitigation strategies. In Spyridon Kintzios (Eds): Aflatoxins: Biochemistry, toxicology, public health, policies and modern methods of analysis. (287- 316), New York, USA, Nova Science Publishers. xi

ORAL COMMUNICATIONS

M Madalena C Sobral, Miguel A Faria, Sara C Cunha, Bojana Miladinovic, Isabel MPLVO Ferreira. Absorption of mycotoxins using in vitro models of human gastric and intestinal epithelium. XXIV Encontro Luso-Galego de Química. Porto, Portugal, 21-23 November 2018.

M Madalena C Sobral, Sara C Cunha, Miguel A Faria, Zita E Martins, Isabel MPLVO Ferreira. Influence of domestic cooking and herbs addition on the exposure to multi- mycotoxins from chicken breast muscle. 4th International Congress on occupational & environmental toxicology. Matosinhos, Portugal, 24-26 October 2018.

POSTER COMMUNICATIONS

M Madalena C Sobral, Roberto Romero-Gonzalez, Miguel A Faria, Sara C Cunha, Isabel MPLVO Ferreira, Antonia Garrido-Frenich. Veterinary drugs stability upon cooking and in vitro digestion of chicken meat burgers. 3rd International Conference on Food Contaminants. Aveiro, Portugal, 26-27 September 2019.

M Madalena C Sobral, Susana Casal, Miguel A Faria, Sara C Cunha, Isabel MPLVO Ferreira. Influence of culinary practices on lipid and protein oxidation in cooked chicken burgers. XX EuroFoodChem. Porto, Portugal, 17-19 June 2019.

M Madalena C Sobral, Sara C Cunha, Miguel A Faria, Zita E Martins, Isabel MPLVO Ferreira. Impact of culinary practices on aflatoxins bioaccessibility in chicken meat. 9th International Congress on Food Digestion. Granada, Spain, 2-4 April 2019.

M Madalena C Sobral, Miguel A Faria, Sara C Cunha, Isabel MPLVO Ferreira. In vitro toxicity of isolated and combined mycotoxins on Caco-2 and HepG2 cell lines. 4th International Congress on Occupational & Environmental Toxicology. Matosinhos, Portugal, 24-26 October 2018.

xiii

TABLE OF CONTENTS

AGRADECIMENTOS ...... V ABSTRACT ...... VII RESUMO ...... IX LIST OF PUBLICATIONS ...... XI LIST OF FIGURES ...... XXI LIST OF TABLES ...... XXV LIST OF ABBREVIATIONS ...... XXVII GENERAL SCOPE AND AIM OF WORK ...... XXXIII THESIS OUTLINE ...... XXXIX CHAPTER I – INTRODUCTION ...... 1 I.1 MEAT CONSUMPTION ...... 3 I.1.1 NUTRITIONAL FEATURES ...... 4 I.1.2 RESIDUAL CONTAMINATION...... 5 I.1.2.1 ANTIBIOTICS ...... 6

I.1.2.2 MYCOTOXINS ...... 7

I.2 DOMESTIC COOKING OF MEAT ...... 9 I.2.1 NUTRIENTS STABILITY DURING COOKING ...... 11 I.2.1.1 PROTEINS ...... 11

I.2.1.2 LIPIDS ...... 13

I.2.1.3 VITAMINS AND MINERALS ...... 15

I.2.2 CONTAMINANTS STABILITY TO COOKING ...... 16 I.2.2.1 FOOD PROCESSING CONTAMINANTS ...... 16

I.2.2.2 ANTIBIOTICS ...... 18

I.2.2.2.1 TETRACYCLINES ...... 18

I.2.2.2.2 Β-LACTAMS AND MACROLIDES ...... 19

I.2.2.2.3 QUINOLONES ...... 20

I.2.2.2.4 SULFONAMIDES ...... 22

I.2.2.2.5 NITROFURANS ...... 22

I.2.2.3 MYCOTOXINS ...... 23

I.3 EXPOSURE TO DIETARY CHEMICAL HAZARDS ...... 24

I.3.1 BIOACCESSIBILITY ...... 24 I.3.1.1 IN VITRO DIGESTION MODELS ...... 26

I.3.2 ABSORPTION ALONG GASTROINTESTINAL TRACT ...... 28 I.3.2.1 IN VITRO ABSORPTION MODELS ...... 30

I.3.3 TOXICITY ...... 33 I.3.3.1 CELL-BASED MODELS TO EVALUATE TOXICOLOGICAL INTERACTIONS ...... 35

I.4 MITIGATION STRATEGIES TO REDUCE EXPOSURE TO DIETARY CHEMICAL HAZARDS 37 I.4.1 COOKING-INDUCED HAZARDS ...... 37 I.4.2 ANTIBIOTICS ...... 39 I.4.3 MYCOTOXINS ...... 39 I.5 FINAL REMARKS ...... 41 CHAPTER II – MATERIAL AND METHODS ...... 43 II.1 REAGENTS, MATERIALS AND STANDARDS ...... 45 II.2 SAMPLES PREPARATION ...... 45 II.2.1 COOKING OF MEAT ...... 45 II.2.1.1 COOKING-INDUCED HAZARDS ...... 45

II.2.1.2 ANTIBACTERIAL AND COCCIDIOSTATS DRUGS ...... 46

II.2.1.3 MYCOTOXINS ...... 47

II.2.2 IN VITRO DIGESTION ...... 47 II.2.2.1 COOKING-INDUCED HAZARDS ...... 48

II.2.2.2 ANTIBIOTICS AND COCCIDIOSTATS DRUGS, AND MYCOTOXINS ...... 49

II.2.3 CELLS-BASED METHODOLOGIES ...... 49 II.2.3.1 CELL CULTURE ...... 50

II.2.3.2 CYTOTOXIC ASSAYS ...... 50

II.2.3.2.1 MYCOTOXINS CYTOTOXIC ASSAY AS MEASURE OF CONTROL FOR TRANSPORT ASSAY ...... 50

II.2.3.2.2 ISOLATED AND COMBINED CYTOTOXICITY ASSAYS ...... 51

II.2.3.3 IN VITRO TRANS-EPITHELIAL TRANSPORT ...... 53

II.3 CHEMICAL ANALYSES ...... 54 II.3.1 NUTRITIONAL ANALYSIS ...... 54 II.3.1.1 PROXIMATE COMPOSITION ANALYSIS ...... 54

II.3.1.2 FREE AMINO ACID ANALYSIS ...... 54

xvi

II.3.1.3 TOTAL FATTY ACID ANALYSIS ...... 54

II.3.2 COOKING-INDUCED HAZARDS ...... 55 II.3.2.1 MALONDIALDEHYDE ANALYSIS (TBARS) ...... 55

II.3.2.2 4-HYDROXY-2-NONENAL AND HEXANAL ...... 56

II.3.2.3 CARBONYLS ...... 56

II.3.2.4 SCHIFF BASES ...... 57

II.3.3 ANTIBIOTICS AND COCCIDIOSTATS DRUGS ...... 57 II.3.3.1 EXTRACTION FOR RAW AND COOKED MEAT ...... 57

II.3.3.2 EXTRACTION FOR BIOACCESSIBLE FRACTIONS ...... 58

II.3.3.3 CHROMATOGRAPHIC ANALYSES ...... 58

II.3.3.3.1 UHPLC-QQQ-MS/MS ...... 59

II.3.3.3.2 UHPLC-ORBITRAP-MS ...... 61

II.3.4 MYCOTOXINS ...... 63 II.3.4.1 EXTRACTION FOR RAW AND COOKED MEAT, AND HERBS ...... 63

II.3.4.2 EXTRACTION FOR DIGESTED AND TRANSPORTED SAMPLES ...... 63

II.3.4.3 CHROMATOGRAPHIC ANALYSIS ...... 64

II.3.5 QUALITY CONTROL/QUALITY ASSURANCE ...... 65 II.3.5.1 QUALITY CONTROL OF CHROMATOGRAPHIC METHODS ...... 65

II.3.5.2 QUALITY CONTROL OF MONOLAYERS FOR TRANSPORT ASSAYS ...... 67

II.4 EXPOSURE ESTIMATION...... 68 II.4.1 BIOACCESSIBILITY ESTIMATION ...... 68 II.4.2 APPARENT PERMEABILITY AND FRACTION ABSORBED ESTIMATION ...... 68 II.4.3 TOXICOLOGICAL EVALUATION ...... 69 II.5 STATISTICAL ANALYSES ...... 71 CHAPTER III – RESULTS AND DISCUSSION ...... 73 III. 1 PROTEIN AND LIPIDS STABILITY ...... 75 III.1.1 IMPACT OF CULINARY PRACTICES ...... 75 III.1.1.1 LIPID OXIDATION ...... 75

III.1.1.2 PROTEIN OXIDATION ...... 80

III.1.2 LIPID AND PROTEIN OXIDATION THROUGH IN VITRO DIGESTION ...... 84 III.1.3 FINAL REMARKS ...... 89

xvii

III.2 ANTIBACTERIAL AND COCCIDIOSTAT DRUGS ...... 90 III.2.1 METHOD DEVELOPMENT AND VALIDATION ...... 90 III.2.2 IMPACT OF CULINARY PRACTICES ...... 94 III.2.2.1 AMOXICILLIN ...... 94

III.2.2.2 FLUOROQUINOLONES ...... 97

III.2.2.3 TETRACYCLINES ...... 98

III.2.2.4 SULFONAMIDES ...... 102

III.2.2.5 MACROLIDES ...... 104

III.2.2.6 COCCIODIOSTATS ...... 104

III.2.3 ANTIBACTERIAL AND COCCIDIOSTAT DRUGS BEHAVIOUR TOWARD IN VITRO DIGESTION ...... 106 III.2.4 EXPLORATORY AND STATISTICAL ANALYSIS ...... 109 III.2.5 FINAL REMARKS ...... 111 III.3 MYCOTOXINS ...... 113 III.3.1 METHOD DEVELOPMENT AND VALIDATION ...... 113 III.3.2 IMPACT OF CULINARY PRACTICES ...... 118 III.3.2.1 EXPLORATORY AND STATISTICAL ANALYSIS ...... 122

III.3.3 MYCOTOXINS’ BEHAVIOUR THROUGH GASTROINTESTINAL TRACT ...... 125 III.3.3.1 EXPLORATORY AND STATISTICAL ANALYSIS ...... 129

III.3.4 MYCOTOXINS GASTRIC AND INTESTINAL ABSORPTION ...... 133 III.3.4.1 QUALITY CONTROL OF MONOLAYERS ...... 133

III.3.4.2 TRANSPORT STUDIES ...... 134

III.3.5 TOXICOLOGICAL INTERACTIONS BETWEEN MYCOTOXINS ...... 146 III.3.5.1 INDIVIDUAL CYTOTOXICITY OF MYCOTOXINS ...... 146

III.3.5.2 CYTOTOXICITY OF BINARY COMBINATION OF MYCOTOXINS ...... 149

III.3.6. FINAL REMARKS ...... 158 III.3.6.1 IMPACT OF COOKING AND IN VITRO DIGESTION...... 158

III.3.6.2 TRANSPORT ACROSS GASTROINTESTINAL EPITHELIUM ...... 158

III.3.6.3 INTESTINAL AND HEPATIC TOXICOLOGICAL INTERACTIONS...... 159

CHAPTER IV – GENERAL DISCUSSION ...... 161 CHAPTER V – CONCLUSIONS AND FUTURE PROSPECTS ...... 167 REFERENCES ...... 171 xviii

ANNEX - SUPPLEMENTARY MATERIAL ...... 197 APPENDIX I ...... 198 APPENDIX II ...... 208

xix

LIST OF FIGURES

Figure I – Schematic overview of the thesis and publications. xxxvii

CHAPTER I - INTRODUCTION

Figure I.1 – Meat supply worldwide (pork, poultry and bovine) between 1995 and 3 2013, according to FAO statistics

Figure I.2 – Schematic representation of a cell monolayer (e.g. Caco-2 cells) 32 grown on a transwell system

CHAPTER III – RESULTS AND DISCUSSION

Figure III.1 – Effect of adding oregano or beer on the TBARS (expressed as MDA) 76 (A), hexanal (HEX) (B), and 4-hydroxy-2-nonenal (HNE) (C) values before and after cooking (oven or microwave) chicken burgers.

Figure III.2 – The variation of ω-3 and ω-6 fatty acids content in oven cooked 77 samples using the raw samples as starting point content (raw content of each condition was subtracted to cooked samples) to evaluate their stability to cooking.

Figure III.3 – Effect of adding oregano or beer on the carbonyl content before and 80 after cooking (oven or microwave) chicken burgers.

Figure III.4 – The variation of free amino acids content in oven cooked samples 81 using the raw samples as starting point content (raw content of each condition was subtracted to cooked samples) in order to investigate the stability of amino acids considered as susceptible to oxidation: (A) Polar, non-charged amino acids; (B) Aromatic amino acids, (C) Apolar amino acids, and (D) Polar, positively charged.

Figure III.5 – Influence of adding oregano or beer on the emission fluorescence 82 spectra (excitation at 360 nm) of raw (A) and cooked samples (B and C). Figures III.5B and III.5C show the variation of fluorescence in cooked samples using the fluorescence of raw samples as starting point content (fluorescence of raw samples of each condition was subtracted to cooked samples) in order to investigate the formation of SB during cooking.

Figure III.6 - Effect of adding oregano or beer on the TBARS (expressed as MDA), 84 hexanal (HEX), 4-hydroxy-2-nonenal (HNE), and carbonyls values after in vitro digestion of oven cooked chicken burgers. The red dashed lines represent the average content of each compound before digestion (content found in cooked samples).

Figure III.7 – The variation of ω-3 and ω-6 fatty acids content in in vitro digested 86 samples (oven cooked samples) using the raw samples as starting point content (raw content of each condition was subtracted to in vitro digested samples) to evaluate their stability to cooking.

xxi

Figure III.8 – Variation of fluorescence in in vitro digested samples using the 87 fluorescence of raw samples as starting point content (fluorescence of raw samples of each condition was subtracted to cooked samples) in order to investigate the formation of SB during digestion with and without ingredients.

Figure III.9– The variation of free amino acids content in in vitro digested samples 88 (oven cooked samples) using the raw samples as starting point content (raw content of each condition was subtracted to digested samples) in order to investigate the stability of amino acids considered as susceptible to oxidation: (A) Polar, non-charged amino acids; (B) Aromatic amino acids, (C) Apolar amino acids, and (D) Polar, positively charged.

Figure III.10 - The impact of oven cooking (A) and microwaving (B) on amoxicillin. 94 Data is expressed as the percentage (%) of ACD that remained after cooking with the addition or not of herbs and/or beer using two initial levels of fortification: 0.1 mg/kg and 1 mg/kg. The dashed line relates to the initial concentration quantified in raw samples, normalized to 100%.

Figure III.11 - Typical ion chromatograms (m/z= 366.11209) of AMX (RT=5.46 min) 96 from HPLC-Orbitrap-MS analysis made in raw, cooked, and digested samples showing the isomerization of AMX into another product sharing the same m/z but different retention time (RT=7.31 min).

Figure III.12 - The impact of oven cooking (A) and microwaving (B) on 97 fluoroquinolones. Data is expressed as the percentage (%) of ACD that remained after cooking with the addition or not of herbs and/or beer using two initial levels of fortification: 0.1 mg/kg and 1 mg/kg. The dashed line relates to the initial concentration quantified in raw samples, normalized to 100%.

Figure III.13 - The impact of oven cooking (A) and microwaving (B) on 99 tetracyclines. Data is expressed as the percentage (%) of ACD that remained after cooking with the addition or not of herbs and/or beer using two initial levels of fortification: 0.1 mg/kg and 1 mg/kg. CTC – chlortetracycline, DC – doxycycline; OTC – oxytetracycline; TC – tetracycline The dashed line relates to the initial concentration quantified in raw samples, normalized to 100%.

Figure III.14 - Typical ion chromatograms (m/z= 479.12201) of CTC (RT=7.75 min) 101 from HPLC-Orbitrap-MS analysis made in raw, cooked, and digested samples showing the isomerization/epimerization of CTC into another products sharing the same m/z but different retention time (RT=6.80-7.40 min).

Figure III.15 - The impact of oven cooking (A) and microwaving (B) on 102 sulfonamides. Data is expressed as the percentage (%) of ACD that remained after cooking with the addition or not of herbs and/or beer using two initial levels of fortification: 0.1 mg/kg and 1 mg/kg. SDZ – sulfadiazine; SDMX – sulfadimethoxine; SMX – sulfamethoxazole; TMP – trimethroprim. The dashed line relates to the initial concentration quantified in raw samples, normalized to 100%.

Figure III.16 - The impact of oven cooking (A) and microwaving (B) on tylosin. Data 104 is expressed as the percentage (%) of ACD that remained after cooking with the addition or not of herbs and/or beer using two initial levels of fortification: 0.1 mg/kg and 1 mg/kg. The dashed line relates to the initial concentration quantified in raw samples, normalized to 100%.

xxii

Figure III.17 - The impact of oven cooking (A) and microwaving (B) on 105 coccidiostats. Data is expressed as the percentage (%) of ACD that remained after cooking with the addition or not of herbs and/or beer using two initial levels of fortification: 0.1 mg/kg and 1 mg/kg. NAR – Narasin, DNC - 4,4’-dinitrocarbanilide. The dashed line relates to the initial concentration quantified in raw samples, normalized to 100%.

Figure III.18 - Analysis of covariance (ANCOVA). Scatter plot of the correlation 111 models between predicted vs observed % ACD after cooking (A) and predicted vs observed % of ACD bioaccessibility (B).

Figure III.19 – The impact of oven cooking (A) and microwaving (B) on impact 119 mycotoxins. Data is expressed as the percentage (%) of mycotoxin that remained after cooking in the presence or absence of herbs and/or roasting bag. The dashed line relates to the initial concentration quantified in raw samples, normalized to 100%.

Figure III.20 – A - Cluster heatmap representation of all eight cooking methods 123 concerning mycotoxins’ percentage after cooking.

Figure III.21 – Cooking impact, translated as estimated standard regression 125 coefficient, for the selected AFB1 (1), AFB2 (2), AFG1 (3), and AFG2 (4), modelled by PLS with 95 % confidence interval (Variable importance for the projection (VIP) >1 are represented in red bold and moderately 0.8< VIP <1 are in black bold).

Figure III.22 – A - Cluster heatmap representation of all eight cooking methods 129 concerning mycotoxins’ bioaccessibility.

Figure III.23 – Bioaccessibility, translated as estimated standard regression 132 coefficient, for the selected AFB1 (1), AFB2 (2), AFG1 (3), and AFG2 (4), modelled by PLS with 95 % confidence interval (Variable importance for the projection (VIP) >1 are represented in red bold and moderately 0.8

Figure III.24 – Transepithelial electrical resistance (TEER) values (Ω cm2) of DON, 134 AFB1, FB1, OTA, and their mixture (MIX) before (0h) and after (3h) the transport experiment for: NCI-N87 cells in the apical à basolateral direction (A) and in the basolateral à apical direction (B); and Caco-2 cells in the apical à basolateral direction (C) and in the basolateral à apical direction (D). Figures III.24E and III.24F show the % of viability of proliferating NCI-N87 and Caco-2 cells, respectively, after 3h exposure to mycotoxins isolated and in mixture.

Figure III.25 – The percentage of DON, AFB1, OTA, and FB1 transferred to the 135 receiver compartment over 180 min across monolayers of NCI-N87 in the apicalàbasolateral and basolateralàapical directions when transported isolated (green dots and lines) or in mixture (orange dots and lines). The dots with SD represent the experimental values obtained in this study, while the line links the theoretical values obtained from the equation showed in Chapter II.4.2.

Figure III.26 – The percentage of DON, AFB1, and OTA transferred to the receiver 136 compartment over 180 min across monolayers of Caco-2 in the apicalàbasolateral and basolateralàapical directions when transported isolated (green dots and lines) or in mixture (orange dots and lines). The dots with SD represent the experimental values obtained in this study, while the line links the theoretical values obtained from the equation shown in Chapter II.4.2. xxiii

Figure III.27 – Apparent permeabilities (Papp) of isolated DON, AFB1, FB1, and 138 OTA transported across NCI-N87 (gastric) and Caco-2 (intestinal) cells over 3h in both AB and BA directions. Grey and white arrows indicate AB or BA transport, respectively, and their thickness mean higher or lower permeability; the red “X” indicates no transport across the monolayer.

Figure III.28 – Fraction absorbed at intestinal level (FA%) of DON, AFB1, and 144 OTA, isolated and in mixture (A); as well as their position in the sigmoidal curve according to their FA% – completely (red), intermediately (orange), and sparingly (yellow) absorbed (B).

Figure III.29 – Percentage of viability (%) of proliferating Caco-2 and HepG2 cells 147 after isolated exposure to serial dilutions of AFB1, DON, FB1 and OTA.

Figure III.30 – Percentage of viability (%) of proliferating Caco-2 and HepG2 cells 150 after 72h of binary combined exposure to AFB1-DON, AFB1-OTA, OTA-DON.

Figure III.31 – Mycotoxins combination plots based on Chou and Talatay 155 combination index theorem after 72 h exposure on Caco-2 cells: (A) combination index plot (CI-Plot) for binary combination of mycotoxins: AFB1-DON, AFB1-OTA, and DON-OTA. (B) Classic isobolograms illustrating the combined toxicity of AFB1-DON, AFB1-OTA, and OTA-DON at IC25, IC50 and IC75.

Figure III.32 – Mycotoxins combination plots based on Chou and Talatay 156 combination index theorem after 72 h exposure on HepG2 cells: (A) combination index plot (CI-Plot) for binary combination of mycotoxins: AFB1-DON, AFB1-OTA, and DON-OTA. (B) Classic isobolograms illustrating the combined toxicity of AFB1-DON, AFB1-OTA, and OTA-DON at IC25, IC50 and IC75.

ANNEX – SUPPLEMENTARY MATERIAL

Figure AI.1 – Texture profile analysis (TPA): A – typical graph obtained showing 207 the first and second bite and the parameters measured; B – Representation of principal component analysis (PCA) and hierarchical cluster analysis (dendrogram) performed to infer which condition better simulates the control condition (F_Ce and F_CI for oven, and M_CE and M_CI for microwaving).

Figure AII.1 – ANCOVA analysis. Means comparison of ACDs percentage after 208 microwave and oven cooking, regardless the other variables studied with a confidence interval of 95%.

xxiv

LIST OF TABLES

CHAPTER I - INTRODUCTION

Table I.1 – Presence of antibiotics in raw meat samples. 7

Table I.2 – Mycotoxins occurrence in meat and meat products. 9

Table I.3 – Cooking methods and their effects on the reduction of antibiotics (%) in 21 meat.

Table I.4 - Overall perception of the impact of common domestic cooking methods 42 on contents of nutrients and contaminants in meat.

CHAPTER II – MATERIAL AND METHODS

Table II.1– Concentrations of the mixtures and concentrations ratios used in the 52 combination experiment assays on Caco-2 and HepG2 cells.

Table II.2 – MS/MS parameters for the UHPLC-MS/MS analysis of antibacterial 60 and coccidiostats drugs.

Table II.3 - UHPLC-Orbitrap-MS parameters for ACDs analysis. 62

Table II.4 - MS/MS parameters for the studied mycotoxins. 65

CHAPTER III – RESULTS AND DISCUSSION

Table III.1 – Proximate composition of raw, oven cooked and microwaved chicken 79 burgers (a), and total fatty acid analysis performed in raw, oven cooked and in vitro digested oven cooked samples (b). All data is expressed as dry weight.

Table III.2 – Recoveries (%), intraday and interday precision (%RSD), limit of 91 quantification (LOQ) of the analytical method applied to determine ACDs in raw meat and digested fraction.

Table III.3 – Percentage of recovery and inter-day precision (%RSD) of ACDs for 93 both dilution method used for bioaccessible samples (n=3), and the method used in non-bioaccessible fraction (NBIO, n=5). The mass balances percentages (%MB) of the in vitro digestion experiment are also presented.

Table III.4 - Percentage (%) of bioaccessibility of ACDs in oven cooked samples 106 (n=6) after duodenal in vitro digestion.

Table III.5 - Percentage (%) of bioaccessibility of ACDs in microwaved samples 107 (n=6) after duodenal in vitro digestion.

Table III.6 - Summary of ANCOVA analysis for cooking and bioaccessibility 110 regarding different cooking methods applied.

xxv

Table III.7 – Recoveries (%), intra-day and inter-day precision (%RSD), limits of 115 detection (LOD) and quantification (LOQ) of the analytical method applied to determine mycotoxins: (A) in the food matrices (cooked chicken breast and herbs); (B) in in vitro digested fractions (gastric and duodenal), (C) after in vitro transport assays (HBSS matrix).

Table III.8 – Multivariate Partial Least Square (PLS) regression between 124 percentage of mycotoxins after cooking (X – variables) and type of cooking applied (i.e. microwaving or oven cooking, presence or absence of bag, and/or herbs addition) (Y – variable).

Table III.9 - Percentage (%) of bioaccessibility of mycotoxins in microwaved 127 chicken breast (n=6) after gastric and duodenal in vitro digestion phases.

Table III.10 - Percentage (%) of bioaccessibility of mycotoxins in oven-cooked 128 chicken breast (n=6) after gastric and duodenal in vitro digestion phases.

Table III.11 – Multivariate Partial Least Square (PLS) regression between 131 percentage of mycotoxins percentage of bioaccessibility (X – variables) and type of cooking applied (i.e. microwaving or oven cooking, presence or absence of bag, and/or herbs addition) (Y – variable).

Table III.12 - Apparent permeabilities (x 10-6 cm/s) and efflux/uptake ratios of DON, 139 AFB1, FB1, and OTA isolated and in mixture in the Apical → Basolateral (AB) and Basolateral → Apical (BA) directions. % MB shows the mass balance recoveries of transport experiments for each mycotoxin.

Table III.13 – Isolated effect parameters calculated from Compusyn software and 148 based on Chou-Talatay method for combination assays.

Table III.14 – Combined effect parameters calculated from Compusyn software and 152 based on Chou-Talatay method for combination assays.

Table III.15 – Dose reduction index values (DRI) at inhibition concentrations of 25, 153 50 and 75% for binary combinations of AFB1, DON and OTA, in Caco-2 and HepG2 cells, that showed to have synergistic effect.

ANNEX – SUPPLEMENTARY MATERIAL

Table AII.1 – Information regarding the formula, molecular weight, purity and 198 supplier of all reagents and specific materials used in this thesis.

Table AII.2 – Information regarding the formula, molecular weight, partition 201 coefficient, and purity and supplier of all standards used in this thesis, as well as the solvent used to prepare the stock and mixed working solutions.

xxvi

LIST OF ABBREVIATIONS

1,3-CHD 1,3-cyclohexanedione 4eATC 4-epianhydrotetracycline 4eTC 4-epitetracycline α-apo-OTC α-apooxytetracycline β-apo-OTC β-apooxytetracycline AB Apical Basolateral direction ABC ATP-binding cassette ACD Antibacterial and coccidiostat drug ACN Acetonitrile AFs Aflatoxins AFB1 Aflatoxin B1 AFB2 Aflatoxin B2 AFG1 Aflatoxin G1 AFG2 Aflatoxin G2 AHC Agglomerative Hierarchical Clustering AHD 1-aminohydantoin AMOZ 3-amino-5-morpholinomethyl-2-oxozolidinone AOAC Association of official analytical chemists AOZ 3-amino-2-oxazolidinone ATC Anhydrotetracycline BA Basolateral  Apical direction BCA Bicinchoninic Acid BCRP Breast cancer resistance protein BHA Butylated hydroxyanisole BHT Butylated hydroxytoluene BSA Bovine serum albumin CI Combination index CIP Ciprofloxacin CLA Conjugated linoleic acid CM Complete medium CTC Chlortetracycline DC Doxycyline DEME Demeclocycline

xxvii

DHA Docosahexaenoic acid Dm Dose of compound required for 50% inhibition DMEM Dulbecco's Modified Eagle Medium DMSO Dimethyl sulfoxide DNPH 2,4-dinitrophenylhydrazine DON Deoxynivalenol DRI Dose reduction index EDTA Ethylenediamine tetraacetic acid EFSA European Food Safety Authority ENR Enrofloxacin EPA Eicosapentaenoic acid ERY Erythromycin ESI Eletrospray ionization EU European Union fa Fraction affected FA Fraction absorbed FAME Fatty acids methyl esters FAO Food and Agriculture Organization FB1 Fumonisin B1 FB2 Fumonisin B2 FB3 Fumonisin B3 FBs Fumonisins FBS Foetal bovine serum FLD Fluorescence detection FLU Flucloxacillin FQs Fluoroquinolones GC-FID Gas chromatography coupled to flame ionization detection GIT Gastrointestinal tract HAAs Heterocyclic aromatic amines HBSS Hank's Balanced Salt Solution HEPES 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid HEX Hexanal HFB1 Hydrolysed fumonisin B1 HHE 4-hydroxy-2-hexenal His Histidine HNE 4-hydroxy-2-nonenal

xxviii

HPLC High performance liquid chromatography HT2 HT-2 toxin IARC International Agency for Research on Cancer ICH International Conference on Harmonization Ile Isoleucine IS Internal standard Leu Leucine LipOx Lipid oxidation LOD Limit of detection LOPs Lipid oxidation products LOQ Limit of quantification Lys Lysine MDA Malondialdehyde MDCK Madin–Darby canine kidney MEM NEAA Minimum Essential Medium Non-Essential Amino Acids MeOH Methanol Met Methionine MIX Mixture MRL Maximum residue limit MRM Multiple reaction monitoring MRP Multidrug resistance protein MS Mass spectrometry MTT Thiazolyl Blue Tetrazolium Bromide MUFAs Monounsaturated fatty acids Mw Microwaving MwB Microwaving with beer MwBg Microwaving with bag MwH Microwaving with herb MwHB Microwaving with herb and beer MwHBg Microwaving with herb and bag NAR Narasin NCZ Nicarbazin NON Nonanal OAT Organic anion transporter OCT Organic cation transporter OFLO Ofloxacin

xxix

OTA Ochratoxin A OTC Oxytetracycline Ov Oven cooking OvB Oven cooking with beer OvBg Oven cooking with bag OvH Oven cooking with beer OvHB Oven cooking with herb and beer OvHBg Oven cooking with herb and bag PAHs Polycyclic Aromatic Hydrocarbons PAMPA Parallel artificial membrane permeability assay Papp Apparent permeability PDCAAs Protein Digestibility-Corrected Amino Acid Scores PG Propyl gallate Phe Phenylalanine PhIP 2-amino-1-methyl-6-phenylimidazo[4,5-b]-pyridine PLS Partial least squares regression P-pg P-glycoprotein ProtOx Protein Oxidation PUFAs Polyunsaturated fatty acids QLs Quinolones QqQ Triple quadrupole mass spectrometer QuEChERS Quick, Easy, Cheap, Effective, Rapid, Safe RMSE Root mean square errors ROB Robenidine ROX Roxithromycin ROS Reactive oxygen species RPMI Roswell Park Memorial Institute RRLC Rapid Resolution Liquid Chromatography SAs Sulfonamides SB Schiff bases SCZ Sulfaclozine SD Standard deviation SDMX Sulfadimethoxine SDZ Sulfadiazine SEM Semicarbazide SFA Saturated fatty acids

xxx

SGF Simulated gastric fluid SIF Simulated intestinal fluid SMX Sulfamethoxazole SMMX Sulfamonomethoxine SMZ Sulfamethoxine SSF Simulated salivary fluid SQX Sulfaquinoxaline T2 T-2 toxin TBA 2-thiobarbituric acid TBARs 2-thiobarbituric acid reactive species TBHQ Tertiary butylhydroquinone TC Tetracycline TCA Trichloroacetic acid TCs Tetracyclines TEER Trans-epithelial electrical resistance TEP Malondialdehyde-bis-(diethyl acetal) Thr Threonine TILM Tilmicosin TJ Tight junctions TMP Trimethoprim TPA Texture profile analysis TPs Transformation products TR True retention Trp Tryptophan Tyr Tyrosine UHPLC Ultrahigh performance liquid chromatography Val Valine WHO World Health Organization ZEN Zearalenone

xxxi

GENERAL SCOPE, AIM OF WORK, AND THESIS OUTLINE

GENERAL SCOPE AND AIM OF WORK

People who want to pursue a balanced diet usually include meat as a relevant source of nutrients for its high-quality protein, minerals and vitamins content. Indeed, compared to other foods, meat offers the advantage of an easier bioavailability of nutrients vital to ensure the body function (1). Therefore, a worldwide increase in meat consumption has been observed in the last decades. Nowadays, the chicken meat has a considerably high relevance in the world concerning its production and nutritional prominence. According to Food and Agriculture Organization of the United Nations (FAO) statistics, an increasing tendency on chicken meat consumption has been observed in the latest years, actually representing the animal livestock with the highest slaughters and animals per area worldwide, but the one with the lowest CO2 production, and therefore lower impact on the environment. Nutritionally, chicken meat has been introduced in society as a healthier alternative diet to red meats, not being associated with cardiovascular disease in epidemiological studies (2). These nutritional health benefits are due to their balanced fatty acid composition (lower ratio of saturated fatty acids (SFAs)/polyunsaturated fatty acids (PUFAs)) in relation to red meat (3). All stages involved in the preparation of a meat-based meal, including animal feeding, slaughter, storage, pre-treatment of meat, and cooking are critical control points for chemical hazards. Thus, chemical hazards from different origins must be considered: i) anthropogenic contamination, namely the antibiotics, concerning the highest livestock production of poultry and their need to prevent diseases; ii) environmental contamination, such as the mycotoxins, due to animals exposure to contaminated feedstuff, and iii) cooking-induced hazards as result of lipid and protein oxidation, for example, PUFAs are more prone to oxidative processes (4-9). The presence of contaminants in meat is inevitable, because the presence of hazardous compounds from oxidation of nutrients and the occurrence of environmental contaminants, like mycotoxins will happen, despite the control. Thus, to avoid their formation as much as possible and to keep the desired nutritional quality of the product mitigation strategies shall be implemented. These strategies are of huge importance to ensure the safety of consumers, because these hazardous compounds may appear at all handling steps of the product, since slaughter, storage, transport to processing in the case of cooking-induced contaminants; and from cropping, storage, transport, and processing in the case of mycotoxins.

xxxiii

GENERAL SCOPE, AIM OF WORK, AND THESIS OUTLINE

Different mitigation strategies can be applied to reduce dietary exposure to contaminants, namely, thermally treatment of meat that can reduce chemical contaminants, but may also favour the formation of cooking-induced ones (10, 11). The use of natural antioxidant ingredients as herbs or marinades may also influence the oxidative processes and avoid the formation of hazardous compounds during cooking (12, 13); in the same way the use of roasting bags has shown to prevent the formation of cooking-induced contaminants while keeping the tenderness of the meat (14). Nevertheless, knowledge concerning the impact of different culinary practices on hazardous compounds already present in raw meat is unknown. Among the three different classes of chemical hazards mentioned above, mycotoxins are the ones whose presence is more difficult to control as they result from environmental contamination, and tend to naturally co-occur in food. Thus, further evaluation on mycotoxins behaviour isolated or in combination with other mycotoxins is of major relevance, taking into consideration that foods co-contamination is frequent and their absorptive patterns across gastrointestinal epithelium may change in the presence of several contaminants, as well as their combined toxic effects which may increase or diminish the toxicity through synergism or antagonism effects.

The primary goal of this thesis was to evaluate the impact of using different culinary practices and ingredients on the stability of food chemical hazards and mitigation strategies during cooking as well as in vitro digestion of chicken meat that can be effective against those chemical hazards. To cover a wide range of chemical hazards likely to be present in that matrix, this thesis investigated the stability of three different classes of chemicals hazards – cooking-induced hazards (lipid and protein oxidation products), anthropogenic hazards (antibiotics), and environmental hazards (mycotoxins), during cooking and digestion processes. As secondary goal, the in-depth evaluation of their interactions during absorption across gastrointestinal epithelium as well as, their combined toxic effect in target organs, such as intestine and liver arose from the ubiquitous presence of mycotoxins in nature and their tendency to co-occur.

To achieve these goals, different specific objectives were established:

1) Undertake an extensive and detailed literature review on the impact of domestic cooking on nutrients and contaminants stability. From this review several gaps were identified: a) Not all cooking methods were evaluated regarding their influence on nutrients and contaminants, and scarce information was available concerning mycotoxins fate during

xxxiv

GENERAL SCOPE, AIM OF WORK, AND THESIS OUTLINE

cooking of meat. Moreover, in the case of nutrients oxidation, the majority of reports were focused on pork and beef, instead of poultry meat. b) Information concerning the fate of some chemical hazards from chicken meat after ingestion was very limited. c) Literature also lacks information concerning the use of culinary ingredients as mitigation strategies to prevent the formation of cooking-induced hazards and/or destroy the hazardous compounds already present in raw meat.

2) Evaluate the impact on lipid and protein oxidation of adding oregano or beer before cooking and in vitro digestion a) Investigate if natural ingredients (oregano or beer) could be used as mitigation strategies to reduce protein and lipid oxidation of meat after oven and microwave cooking; b) Monitor protein and lipid oxidation through in vitro digestion, and infer if herbs or beer are able to diminish hazards formation and keep the nutritional value of meat.

3) Explore culinary practices that can reduce the final exposure to antibiotics from contaminated chicken meat a) Evaluate the impact of oven and microwave cooking allied with the use of oregano and/or beer on antibiotics stability from chicken meat; b) Investigate the behaviour of antibiotics during digestion of contaminated cooked meat with and without the added ingredients.

4) Study the mycotoxins fate during cooking using commercial mixture of herbs and roasting bags and in vitro digestion of contaminated chicken meat a) Understand the stability of mycotoxins from chicken meat during oven and microwave cooking allied with the use of extra-ingredients (commercial mixture of herbs and roasting bags) as mitigation strategies b) Investigate the behaviour of mycotoxins during digestion of contaminated cooked meats

5) Understand the absorption patterns of mycotoxins isolated or in mixture across the gastrointestinal epithelium. a) Evaluate the isolated transport of selected mycotoxins across gastric NCI-N87 and intestinal Caco-2 monolayers. b) Perform the same transport assay using a mixture of the used mycotoxins to assess if a mixed exposure interferes with the absorption pattern of isolated mycotoxins.

xxxv

GENERAL SCOPE, AIM OF WORK, AND THESIS OUTLINE

6) Evaluate the impact of toxicological interactions of mycotoxins on hepatic and intestinal cells. a) Evaluate the isolated toxicity of selected mycotoxins after exposure to proliferating intestinal Caco-2 and hepatic HepG2 cell lines. b) Determine the type of the interaction as synergism, addition or antagonism performing binary combinations of mycotoxins using the same cells lines.

Figure I shows a schematic overview of the thesis and publications in line with thesis specific goals.

xxxvi

GENERAL SCOPE, AIM OF WORK, AND THESIS OUTLINE

Figure I- Schematic overview of the thesis and publications.

xxxvii

GENERAL SCOPE, AIM OF WORK, AND THESIS OUTLINE

THESIS OUTLINE

This thesis includes five chapters, Chapter I presents a detailed literature review related with the topic of thesis: firstly, the overall meat consumption, nutritional features and contamination; then, the cooking methods usually applied to cook meat and their impact on nutrients and contaminants; later, the exposure to dietary chemical hazards, and finally, which mitigation strategies have been used to reduce dietary chemical hazards exposure.

The Chapter II entitled of “material and methods”, summarizes the experimental design of all studies performed during this thesis, together with the materials and methods used.

The Chapter III comprises the results and discussion section, and is divided in three groups corresponding to each chemical hazard studied: cooking-induced hazards (Chapter III.1), antibiotics (Chapter III.2), and mycotoxins (Chapter III.3).

Chapter III.1 is divided in three sections, the first two focus on the impact of using natural ingredients (herbs or beer) against the oxidative processes, namely lipid and protein oxidation of chicken meat during cooking (Chapter III.1.1) and during in vitro digestion (Chapter III.1.2) finishing with a brief conclusion (Chapter III.1.3).

Chapter III.2 is divided in five sections, the first concerning method development and validation (Chapter III.2.1), followed by the impact of using the mitigation strategies to investigate the stability of antibiotics towards cooking (Chapter III.2.2) and in vitro digestion (Chapter III.2.3). Then, the Chapter III.2.4 presents an exploratory and statistical analysis of data, followed by a brief conclusion of the chapter (Chapter III.2.5).

Chapter III.3 is divided in six sections, the first concerning method development and validation (Chapter III.3.1). Then, Chapter III.3.2 relates to the fate of mycotoxins during cooking and Chapter III.3.3 investigates their behaviour towards in vitro digestion. Afterwards, the absorption of mycotoxins, individual or in mixture, across gastric (NCI-N87) and intestinal (Caco-2) monolayers was ascertained (Chapter III.3.4); and finally, Chapter III.3.5 investigates hepatic (HepG2) and intestinal (Caco-2) toxicity of the same mycotoxins isolated or in binary combination. Chapter III.3.6 shows the final remarks of each study.

Chapter IV presents the general discussion of this thesis. xxxix

GENERAL SCOPE, AIM OF WORK, AND THESIS OUTLINE

Chapter V deals with the overall conclusions of this thesis as well as the future prospects.

At the end a list of all the references cited in this thesis is included to support the information herein introduced and discussed as well as the supplementary material of this thesis.

CHAPTER I – INTRODUCTION

I.1 MEAT CONSUMPTION

I.1.1 NUTRITIONAL FEATURES

I.1.2 RESIDUAL CONTAMINATION

I.2 DOMESTIC COOKING OF MEAT

I.2.1 NUTRIENTS STABILITY DURING COOKING

I.2.2 CONTAMINANTS STABILITY TO COOKING

I.3. EXPOSURE TO DIETARY CHEMICAL HAZARDS

I.3.1 BIOACCESSIBILITY

I.3.2 ABSORPTION ALONG GASTROINTESTINAL TRACT

I.3.3 TOXICITY

I.4 MITIGATION STRATEGIES TO REDUCE EXPOSURE TO DIETARY CHEMICAL HAZARDS

I.5 FINAL REMARKS

Parts of this chapter are available in the following publications:

Sobral MMC, Faria MA, Cunha SC, Ferreira IMPLVO. Aflatoxins in food and feed: occurrence, legislation and mitigation strategies. In Spyridon Kintzios (Eds): Aflatoxins: Biochemistry, toxicology, public health, policies and modern methods of analysis. (287-316), New York, USA, Nova Science Publishers.

Sobral MMC, Faria MA, Cunha SC, Miladinovic B, Ferreira IMPLVO. Transport of mycotoxins across human gastric NCI-N87 and intestinal Caco-2 cell models. Food and Chemical Toxicology. 2019. 131:110595.

Sobral MMC, Romero-Gonzalez R, Faria MA, Cunha SC, Ferreira IMPLVO, Garrido-Frenich A. Stability of antibacterial and cocciodiostat drugs on chicken meat burgers upon cooking and in vitro digestion. Food chemistry. 2020. 316:126367.

CHAPTER I. INTRODUCTION

I.1 MEAT CONSUMPTION

Anthropologists believe that regular meat consumption was crucial to the human evolutionary heritage (15, 16). Hominids began using fire to cook foods approximately 1.8 million years ago and this played an important role in the improvement of several human- specific characteristics. The energy gained from cooked food contributed to the development of larger brains in the human lineage, as well as to several adaptations from the mouth to the intestinal tract due to the consumption of cooked foods (15). In the second half of 20th century, a fast global population growth was observed and along with it an expected growth in meat production. Meat production has quadrupled since 1961, nowadays exhibiting a global production exceeding 320 million tonnes each year and representing an average of 43 kg consumption per capita in 2014 (17). These high meat consuming patterns are observed in high-income countries, where meat and meat products are essential components in diet, and their consumption is associated to healthy and prosper meal (3), however, it shall be mentioned that meat production has a large environmental impact, increasing greenhouse gas emissions and using the highest land area for livestock productions in comparison with other foods. According to European legislation, the term meat refers to the edible portions removed from the carcass of domestic animals or animals farmed as domestic animals, like bovine, porcine, ovine, caprine, and poultry (18). According to FAO statistics, the actual dominant livestock meat types are pork, chicken, and bovine meat, with the first two exhibiting an increase over the years, while the last experienced a slight decrease in consumption (Figure I.1).

Figure I.1 – Meat supply worldwide (pork, poultry and bovine) between 1995 and 2013, according to FAO statistics (19). 3

CHAPTER I. INTRODUCTION

In this sense, pork meat with an average per capita consumption of 16 kg represented the highest meat commodity in 2013, followed by poultry meat (~13kg consumption per capita), and bovine meat (~8kg consumption per capita). The decrease of bovine meat consumption is likely related to the negative health effects associated with this type of meat. Red meat is receiving negative evaluations in the past decades due to its possible carcinogenic effects, encouraging a reduction of their consumption. In October of 2015, the International Agency for Research on Cancer (IARC) released a report classifying processed meat as carcinogenic (group 1) and red meat as probably carcinogenic (group 2A) to humans (20). In addition, red meat has been associated with higher incidences of cardiovascular diseases being poultry, by contrast, associated with a healthy dietary alternative to red meats in epidemiological studies (2).

I.1.1 NUTRITIONAL FEATURES

Meat has high nutritional value providing an important source of macronutrients, such as a high-quality proteins and unsaturated fatty acids. It is also a source of micronutrients, essential for growth and development (21, 22), such as minerals, namely selenium, zinc, copper, and heme-iron, the iron form with the highest bioavailability; and vitamins, especially from the B-complex, representing the main source of B12 vitamin in diet. Moreover, a serving of pork meat is able to provide a fulfilment of a daily thiamine requirement and is one of the main sources of niacin (23). Meat is a well-recognized protein source, representing on average 22% of its edible portion, according to the Portuguese nutritional table data (24). Furthermore, meat protein is considered of high nutritional value concerning its high digestibility score of 0.92 out of 1.0, as determined by the Protein Digestibility-Corrected Amino Acid Scores (PDCAAS), a score higher than the protein sources of plant origin (25). Meat proteins are also recognized by their high content of essential amino acids, which are mandatory amino acids to ensure the protein synthesis and function in body (26), and since these cannot be produced by the humans, they must be supplied by the diet (25). Meat nutritional properties are also related to its fat content and fatty acid composition, since it gives flavour to food and helps in the absorption of lipophilic compounds, such as fat soluble vitamins (A, D, E and K) that play an important role in the immune response (27). However, meat is seen as a major source of fat in diet, especially the saturated fatty acids, which have been implicated in cardiovascular diseases. The fatty acid composition of meat shows on average 50% saturated fatty acids (SFA), 40% monounsaturated fatty acids (MUFAs), and 10% of polyunsaturated fatty acids (PUFAs) but it may differ on type of meat,

CHAPTER I. INTRODUCTION origin and feed. The high content of SFA is the reason why red meats have been linked to several negative health effects, namely, cancer or coronary heart diseases (28). By contrast, the several health effects associated with poultry meat consumption are due to its lower fat content and higher proportion of PUFAs compared to other meats (3). The contribution of total fat, SFA, ω-6 PUFAs, ω-3 PUFAs, and trans-fatty acids for the total energy intake should not exceed 15-30%, 10%, 5-8%, 1-2% and 1%, respectively. Thus, while the reduction of SFA intake is encouraged, it is also suggested an increase of ω-3 PUFAs intake, and meat represents an important source of ω-3 PUFAs for humans. In this sense, meat contributes with the intake of important fatty acids such as eicosapentaenoic acid (EPA, 20:5 ω-3) and docosahexaenoic acid (DHA, 22:6 ω-3) that are important to reduce the risk of cardiovascular diseases, critical for proper brain and visual development in foetus and through life. Meat also contains araquidonic acid (20:4, ω-6), which is the precursor of prostaglandins that mediate several reactions in the body in response to inflammation (29). Howe and colleagues (30) stated that for Australian adults 43% of total dietary intake of ω-3 polyunsaturated fatty acids could be achieved by consuming meat products due to their higher meat intake. Moreover, in a daily diet meat contributes with more than 15% of energy intake, 40% of protein intake and 20% of fat intake. Pork and poultry meat together represent 80% of total meat consumption in European countries (31).

I.1.2 RESIDUAL CONTAMINATION

Meat is also prone to contamination, contributing to the human exposure to contaminants, such as, antibiotics, pesticides, heavy metals, among others (4). The increasing demand for the livestock production of terrestrial animals has led to increased occurrence of veterinary drugs in meat destined for human’s consumption, as a result of their use to fight infections in animals (8). Emergent contaminants like mycotoxins, can also enter the food chain via livestock feed contamination, and carry-over to humans via meat, milk or eggs consumption (5-7). In the future, increased mycotoxins outbreaks are expected in food as a result of climate changes (i.e. increase of temperature and moisture that allow moulds to grow). The following sub-sections will give detailed information about raw meat contamination, regarding two classes of chemical hazards: antibiotics and mycotoxins.

CHAPTER I. INTRODUCTION

I.1.2.1 ANTIBIOTICS

Antibacterial drugs are commonly used as therapeutic agents to treat diseases in farmed animals; these drugs are also used prophylactically to prevent animal infections, and as additives in feedstuff production. The overuse of, and improper treatment with, these antibiotics, such as insufficient withdrawal time, may lead to the presence of residues in commercialized meat products. The large scale production and use in several areas, among them the animal production, has contributed with a significant amount of antibiotics into the environment, and antibiotics contamination has been recognized as a global environment issue (32). Additionally, the antimicrobial nature of antibiotics confers resistance to biodegradation contributing to their permanent presence in the environment, and for this reason they gained the classification of pseudo-persistent organic pollutants (33). Sulfonamides and fluoroquinolones are the antibiotics most commonly used for veterinary purposes, showing 20 and 19% frequencies of detection in food samples, respectively; these are followed by β-lactams and tetracyclines, with 15 and 8% frequencies of detection, respectively (34). Antibacterial residues in meat may pose a serious threat to human health by damaging intestinal flora, causing the development of resistant strains, leading to the failure of antibiotic therapy in clinical situations, and cause allergic reactions in hypersensitive individuals (35, 36). Table I.1 shows a summary of studies related to the presence of antibiotics in raw meat samples. Most studies were conducted using poultry rather than other meats. The occurrence of tetracyclines - doxycycline (DC), oxytetracycline (OTC), and chlortetracycline (CTC) - was reported in chicken samples, with OTC exceeding the maximum residue limits (MRL) established by legislation (100 µg/kg) (37). Fluoroquinolones are the second major antibiotic group used as a veterinary drug and their presence is detected in the muscle tissues of chickens (Table I.1). In Nigeria, ciprofloxacin, norfloxacin, and ofloxacin are found in 15 to 55% of muscle tissues from chicken, pork, and chevon, with mean concentrations of 67.85- 345.62, 11.39-173.40, and 12.37-79.28 µg/kg wet weight, respectively (38). Similarly, Pena and colleagues (39) studied the presence of enrofloxacin, ciprofloxacin, norfloxacin, and sarafloxacin in 98 samples of poultry muscle from Portuguese market; 27% of the samples tested positive in the range of 20.9-164.9 µg/kg wet weight. The occurrence of quinolones in meat matrices is described in Table I.1. In some cases, the occurrence of quinolones exceeds the permitted MRL (100 µg/kg). Controlling the use of such antibiotics in livestock production is important and urgently needed to prevent human exposure via consumption of contaminated animal-derived products.

CHAPTER I. INTRODUCTION

Table I.1 – Presence of antibiotics in raw meat samples. Concentration Matrix Ref (range, µg/kg, ww) TETRACYCLINES OTC Chicken 12.7-1146.1 (40, 41) CTC Chicken 10.0-11.0 (40) DC Chicken Nd – 12.0 (40) QUINOLONES ENR Chicken 2.8-333.2 (39, 42, 43) Chicken, Chevon, CIP 67.9-345.6 (38) Pork, Beef SULFONAMIDES SDZ Chicken 8.0-62.0 (44) SMZ Chicken 4.0-41.0 (44) SMX Chicken 5.0-11.0 (44) SQX Chicken 2.0-8.0 (44) SCZ Chicken 93-2710 (43) OTC – oxytetracycline; DC – doxycyline; CTC – chlortetracycline; ENR – enrofloxacin; CIP – ciprofloxacin; SDZ – sulfadiazine; SMZ – sulfamethoxine; SQX – sulfaquinoxaline; SCZ – sulfaclozine; Nd – not detected; ww – wet weight;

I.1.2.2 MYCOTOXINS

Mycotoxins are toxic secondary metabolites from fungi that may cause health problems to humans and animals upon ingestion, inhalation or skin contact (5). More than 300 fungal metabolites are produced by Aspergillus, Penicillium and Fusarium fungi species and recognized as mycotoxins (45, 46). The most prevalent foodborne fungal toxins are aflatoxins B1, B2, G1, G2, and M1, thrichothecenes (deoxynivalenol (DON), HT-2 toxin (HT2) and T-2 toxin (T2)), fumonisins B1, B2 and B3, ochratoxin A (OTA), and zearalenone (ZEN). Mycotoxins occurrence in food are an important topic regarding the actual human’s health risk assessment considering the climate changes (i.e. increase of temperature and moisture that benefit moulds to grow) that will likely contribute with an increase of mycotoxins outbreaks in food. Moreover, recent data shows that these metabolites are likely to naturally co-occur in food and feedstuff (47). As reported by Rodrigues and Naehrer (48), 48% of the analysed samples were contaminated with 2 or more mycotoxins, which turns the human co-exposure to multiple mycotoxins a real problem, increasing the concern about their combined impact on health. Mycotoxins can enter the food chain via fungal infection of crops, which contribute to livestock feed contamination and consequent animal exposure through contaminated feedstuff. Animal ingestion of mycotoxins can result in their accumulation in various organs and tissues, contributing for humans’ exposure via meat, milk or eggs consumption (7). Thus, mycotoxins presence in meat products usually results from indirect transmission from animals exposed to naturally contaminated feed (49). Mycotoxins are likely to undergo biotransformation in liver resulting in the formation of other species of mycotoxins (e.g. aflatoxin M1 and M2, hydrolysed fumonisin B1 (HFB1), ochratoxin-α, α- and β-zearalenol,

CHAPTER I. INTRODUCTION among others) (50), however, the presence of aflatoxin B1, fumonisins B1 and B2, OTA and ZEN was reported in animal feed and meat (beef, sausage, pork, and chicken) suggesting that mycotoxins contamination in feedstuff directly influences their presence in food products of animal origin (7, 51-54). Trace amounts of DON and T2 were equally reported in pork fat, with the later one being also detected in pork muscle and chicken muscle (55) (Table I.2). Beyond the abovementioned contamination, these metabolites can also reach meat and meat products through the growth of toxigenic moulds on the surface of cured and ripened meat products (56). Due to their ubiquity and tolerance to diverse environmental conditions – pH, water activity, NaCl concentration, and storage conditions (e.g. temperature, humidity, and oxygen availability) –, several microbial groups are capable of growing in meat and meat products. Penicillium spp. and Aspergillus spp. are the most commonly isolated moulds from sausages and cured ham (56). Moreover, as meat quality is dependent on diet, the administration of contaminated feed to animals will affect the characteristics of meat because mycotoxins influence the metabolism of animals (57). Previous reports showed that mycotoxins contamination negatively affects organoleptic characteristics (e.g. colour) of chicken meat, and in specific, T2, DON and FBs induced high lipid peroxidation in meat (58, 59). Strict regulations have been established to control mycotoxins occurrence in feedstuff and cereal/spices intended for direct humans’ consumption (60, 61). The use of preventive strategies, among them the physical treatment, to control mould growth and degrade mycotoxins, were also taken into account to set those limits. Nevertheless, maximum levels are not regulated for products from animal origin like meat and meat products that are also prone to mycotoxin contamination.

CHAPTER I. INTRODUCTION

Table I.2 – Mycotoxins occurrence in meat and meat products. Concentration Mycotoxins Matrix Ref (range, µg/kg) Aflatoxins Total AFs Chicken muscle* Nd-4.90 (53) Beef luncheon 0.47-2.1 (51) Beef burger 0.55-7.5 (51) AFB1 Chicken muscle* Nd-4.41 (52, 53) Swine muscle 0.46-0.74 (52) Ham Nd-1.06 (7) Sausages Nd-1.69 (7, 54, 56) Sucuk** 0.06-0.129 (56) Pastirma*** 0.049-47 (56, 62) AFB2 Sucuk** 0.027-0.067 (56) Sausage 0.020 (56) AFG1 Sucuk** 0.077 (56)

Tricothecenes DON Pork back fat 0.1232-0.4265 (55) T2 Pork dorsal muscle 0.024-0.4515 (55) Pork back fat 0.024-0.0906 (55) Chicken muscle 0.0704-0.0904 (55)

Fumonisins FB1 Sausages 22-27 (54) FB2 Sausages 16-36 (54)

Ochratoxins OTA Beef luncheon 0.56-8.5 (51) Beef burger 2.7-7.6 (51) Beef 0.01-0.03 (49) Ham 0.98-9.95 (7) Sausages 0.95-3.13 (7, 48) Bacon Nd-1.23 (7) Pork 0.01-0.88 (49, 52) Zearaleone ZEN Sausages 10-23 (54) Pork 4-10 (54) Nd – non detected; AFs – aflatoxins; AFB1 – Aflatoxin B1, AFB2 – Aflatoxin B2; AFG1 – Aflatoxin G1; DON – deoxynivalenol; T2 – T-2 toxin; FB1 – fumonisin B1; FB2 – fumonisin B2, OTA – ochratoxin A; ZEN – zearalenone; *wings, chest and legs; **dry, spicy sausage eaten from the Balkans (Middle East and central Asia);***highly seasoned, air-dried cured beef.

I.2 DOMESTIC COOKING OF MEAT

Cooking meat in modern times can be conducted by different methods, including boiling, steaming, roasting, baking, frying, microwaving, grilling, barbecuing, smoking, sous-vide and confit. Boiling is a simple cooking method widely used to cook meat, involving water at the boiling point of 100 °C. Steaming relies on cooking with steam heat resulting from boiling water. Meat is maintained separate from the boiling water, having direct contact only with steam; this contributes to the moist texture of meats cooked with steam. Oven roasting is a common cooking method used to enhance the flavour of meat via caramelization and Maillard browning occurring on the surface of the food; this is accomplished using dry heat, at temperatures above 200 °C, over variable cooking times. Similarly, oven baking is a cooking method that uses dry heat at lower temperatures (170 °C) than those used for oven 9

CHAPTER I. INTRODUCTION roasting and, usually, the food is covered while cooking (63). Frying is a cooking method in which food is submerged in hot fat, most commonly in cooking oil. Typically, it is a rapid preparation technique that promotes physical and chemical changes in the products and leads to unique sensory properties of colour, flavour, texture, and palatability (64). There are different types of frying, the most common being pan- and deep-frying. The first consists of cooking the food in a hot pan containing a scant amount of cooking oil; in the second one, the pan contains enough cooking oil to allow the food to float in the cooking oil. Microwaving is a more recent cooking method, widely applied to cook and reheat meals. Microwave ovens work by producing electromagnetic waves via a magnetron contained inside the oven. These waves cause the vibration of water molecules in the food, producing heat that is then transferred throughout the product by thermal conduction. One of the major problems associated with microwave heating is non-uniform temperature distribution. This can result in cold spots, which may pose serious health concerns with respect to meat products, because of the survival of bacteria in the cold spot of the meat (65). Recent studies have focused on how microwave reheating affects nutrients. For example, cholesterol appears to undergo higher oxidation when reheated in a microwave oven compared to that using other reheating methods (11, 66). Grilling allows consumers to prepare a quick meal by using a direct heat source; the heat source, such as thermal radiation or direct conduction, may differ according to the type of grill. Direct grilling can expose food to a temperature up to 260 °C, resulting in grilled meat with aroma and flavour characteristics similar to those achieved by roasting; this is due to the Maillard reaction. Contrary to grilling, barbecuing is a slow cooking process, using the smoke produced by burning wood or charcoal as indirect heat source, resulting in smoking-flavoured, tender and juicy meat. Smoking is an ancestral flavouring and preserving method commonly used with meat. Cold-smoking is used to enhance the flavour of chicken breasts, beef, pork, salmon, and steak; some of these products are then baked, grilled, or roasted before eating. Cold- smoking temperatures normally range between 12-25 °C. Hot smoking, which exposes food to smoke and heat in a controlled environment, occurs at 40-100 °C, with the internal temperature of the smoked product reaching 85 °C (67). Hot-smoked foods are safe to eat without further treatment, although they are often reheated or cooked. Sous-vide is a French expression for “under vacuum”. It consists of vacuum-sealing meat in a heat-stable airtight plastic bag, placed in a water bath, or in a temperature- controlled steam environment, for an extended cooking time (hours) at low temperature (<100 °C). Sous-vide prevents the loss of aromatic volatiles and preserves moisture content, resulting in especially flavourful and nutritious food (68). Similarly, confit consists

CHAPTER I. INTRODUCTION of cooking meat in a fat source at low temperatures (< 90 ºC) for a long time (3-13h). This cooking method is often applied to cook duck, pork, and goose meat.

I.2.1 NUTRIENTS STABILITY DURING COOKING

Cooking increases the safety of food by inactivate the growth of microorganisms, inactivating anti-nutrient enzymes, and increasing shelf-life. Additionally, it improves the organoleptic properties of meat (69, 70). Cooking with heat denatures proteins and increases the digestibility and bioavailability of nutrients (71, 72). The temperature and time of cooking play an important role in the characteristics of cooked meat. Myofibrillar and connective tissue proteins (collagen and elastin) control the toughness of muscle tissues. During heating, these proteins are denatured, causing destruction of cell membranes, shrinkage of fibres, aggregation and gell7ing of myofibrillar and sarcoplasmic proteins, and shrinkage and solubilisation of connective tissue (73, 74). Thus, cooked meat is easier to chew, digest, and is, therefore, more nutritious than uncooked meat. This is caused by the unwinding of proteins and loss of connections between muscle fibres, facilitating digestion in the small intestine. However, cooking may affect the nutritional composition of meat, including water content, fat profile, amino acids, and bioactive compounds such as vitamins and polyphenols, influencing the final quality of the product (23, 75-79). The effects of boiling, baking, grilling, frying, and microwaving on the proximate and mineral composition of meat have been described in the literature (80-83). Generally, cooking contributes to the loss of water-holding capacity, resulting in the concentration of proteins, fat, and ash in cooked meat (81, 83, 84). However, specific changes, resulting from the cooking process, are very diverse; thus, a brief discussion, concerning the impact of domestic cooking on the content of proteins, lipids, vitamins, and minerals in meat, is presented in the following sub- sections.

I.2.1.1 PROTEINS

Cooking is linked with the formation of reactive oxygen species (ROS), which may contribute to the oxidation of nutrients (85). The protein oxidation, observed during cooking occurs via: i) increased surface hydrophobicity caused by unfolding of proteins via the loss of hydrogen bonds and subsequent exposure of hydrophobic amino acids (86, 87); ii) aggregation of meat proteins due to covalent modifications, caused by the formation of disulfide and dityrosine bridges, resulting in loss of thiol groups and oxidation of tyrosine (86, 88, 89); and iii) increase of carbonylation, which is caused by conversion of basic amino 11

CHAPTER I. INTRODUCTION acids, such as lysine, histidine, and arginine, into carbonyl derivatives via attack of ROS, as well as carboxylation and Schiff bases formation (87, 89, 90). The latter is a precursor in Maillard reactions, which are responsible for the browning of meat during cooking. These processes result in losses of essential amino acids leading to an overall decrease of protein quality. Regarding the 20 common amino acids side chain and the peptide backbone, a large number of different radicals can be generated on proteins, but also with free amino acid although at lesser extent. In in vivo situations, the electrophilic radicals are more common than nucleophilic radicals, with the first ones (e.g hydroxyl (HO•), alkoxyl (RO•), and peroxyl (ROO•)) preferentially oxidizing electron-rich sites (91). The amino acids more prone to oxidation are the aromatic amino acids such as phenylalanine, tyrosine, and tryptophan; the amino acids containing an amine group in the side chain like arginine, histidine, and lysine; and the hydrophobic amino acids, namely valine, leucine, and isoleucine. The later ones are susceptible to oxidation because protein hydrophobic residues are usually in the core of the proteins and when these are denatured the amino acids are exposed to unfavourable conditions (91). Protein oxidation has already been studied in food focusing on its evolution during storage (92) and more recently concerning the impact of thermal treatment on its formation (87, 93). The influence of heat treatment/cooking on meat proteins has been reviewed by Yu and colleagues (74); and the implications of meat protein oxidation on human health have been addressed by Soladoye and colleagues (85). The extent of protein oxidation highly depends on time and temperature of cooking. Cooking beef (M. Rectus abdominis) and pork (M. Longissimus thoracis) at 100 ºC for 5 to 30 min increased protein hydrophobicity and protein aggregation (86, 87). These oxidation products were correlated with drip loss, suggesting the loss of water holding capacity of oxidized proteins after cooking. Protein oxidation can also be evaluated by increase of carbonyl contents and Schiff bases formation. A two fold increase in carbonyl contents was observed after cooking beef at 100 ºC for 5 min compared to raw sample, thus increasing with time. Cooking at high temperatures (270 ºC) for 1 min highly increased carbonyl contents at the same extent as cooking at 100 ºC for 45 min (86). Such rapid protein oxidation can be attributed to the loss of antioxidant protection of muscles as glutathione peroxidase and catalase activity that drastically drop after heat treatment. Cooking M. Longissimus thoracis at 100 ºC (10 or 30 min) or at extreme conditions (207 ºC, 300 s) strongly increases carbonyl contents and Schiff bases formation, compared with raw meat; while cooking at 65 and 96 ºC results in fewer carbonyl formation in cooked beef (87, 89). Sous-vide cooking of lamb loins (60 or 80

CHAPTER I. INTRODUCTION

ºC for 6 or 24h) increases carbonyl contents at all cooking temperatures reaching levels 3.5 times higher than raw meat (90). The oxidative reactions, which contribute to protein oxidation, may result in the loss of amino acids in cooked meat compared with the content of raw samples. Cooking pork at 60 and 75 °C decreased the protein content by 11 and 18%, respectively; histidine is the most affected amino acid, undergoing a reduction by 17 and 31%, respectively (79). The same authors suggested that, considering the actual loss of amino acids, more than 130 g of cooked pork meat should be consumed instead of the recommended 100 g to achieve the same amino acid intakes. Cooking temperature was the determining factor with respect to the stability of aromatic amino acid residues. Cooking bovine meat at 60 °C did not affect the stability of amino acids, but their contents were highly reduced when cooked at 100 and 140 °C. Their thermal stability was ranked as tryptophan > phenylalanine > tyrosine (94). Cooking pork in the oven at 102 °C for 20 min shows a 50% reduction in the content of tryptophan (95). Grilling, microwaving, and smoking the PDO (Protected Denomination of Origin) veal breed “Barrosã” meat significantly increases the content of several amino acids; this occurs because of water loss. Leucine, glutamic, and aspartic acids are the predominant amino acids present, whereas cysteine, asparagine, glutamine, and tryptophan were not detected in cooked meat due to their destruction by thermal hydrolysis (78). In this sense, proteins are oxidized during cooking which may lead to the loss of nutritional value of meat. Not all cooking methods have been evaluated concerning their impact on protein neither mitigation strategies have been proposed to control/avoid oxidative processes.

I.2.1.2 LIPIDS

Several cooking processes, such as boiling, microwaving, grilling, and frying, can affect the lipid composition and nutritional value of cooked products compared with those in the raw samples (75, 77, 96). The type of meat, content of fats, and level of doneness highly influences the fatty acid composition, as regard of saturated or unsaturated fatty acids, in cooked meat (77). Roasting (200 ºC, 12 min), grilling (130-150 ºC, 5 min), microwaving (1000 W, 1.5 min), and frying (170-180 ºC, 4 min) foal steaks increased fat content values, from 0.33% (raw meat) to 1.15, 1.21, 1.46, and 2.54%, respectively. All cooking methods decreased SFA contents of meat, and in the case of frying, the SFA contents were replaced by MUFAs derived from oil used during cooking. From a nutritional point of view, roasting and grilling are suggested as the best cooking methods to cook foal meat (76). The major

CHAPTER I. INTRODUCTION changes in the composition of fatty acids, verified by Alfaia and colleagues (75), include higher percentages of SFAs and MUFAs, and lower percentages of PUFAs, in cooked beef compared with those in raw meat; however, conjugated linoleic acid (CLA) isomers show high stability during thermal processing. The nutritional value of fat for human consumption is widely evaluated by the ratio of PUFA/SFA that must be higher than 0.45, and, within PUFAs content, the ω-6/ω-3 PUFAs ratio that must not exceed 4. Frying is one of the cooking methods responsible for the major changes in fatty acid composition due to the exchange lipid content with oil used during cooking. PUFA/SFA ratios of fried foal were higher (0.94) than fried beef (0.42-0.65) and pork (0.13-0.18). The ω-6/ω-3 ratio of foal meat (1.02–1.39) seems not to be affected by frying, while frying beef highly increases this ratio to 5.91-6.47 (75-77). Thus, in the case of frying meat, foal meat seems to be nutritionally better than pork and beef regarding fatty acid composition. The aroma and flavour development in fried foods is a complex process mainly derived from lipid oxidation and Maillard reactions (64, 97). The presence of PUFAs renders the product more susceptible to oxidation during heating. Lipid oxidation is highly dependent on the origin of meat, type of muscle, species, and storage conditions with the susceptibility of meats to undergo oxidation processes (98). Also, these oxidative reactions rely either on endogenous factors such as heme-iron content and the natural antioxidant mechanisms of meat (that partially collapse after slaughter), and on external factors such as temperature, oxygen availability, and pH during handling and storage (98). Poultry meat is highly sensitive to oxidation owing to the unsaturation degree of their lipids (99). PUFAs like linoleic (18:2, ω-6) or arachidonic acids (20:4, ω-6) are susceptible to free radical attack giving rise to primary products such as hydroperoxides which readily decompose to secondary products as ketones, alcohols, and aldehydes (100). Malondialdehyde (MDA) and 4- hydroxy-2-nonenal (HNE) are well-recognized reactive aldehydes derived from oxidation of fatty acids, known to form adducts with DNA and proteins. These reactions lead to the alteration in their functions and cause health concerns, such as metabolic and neurodegenerative diseases, and cancer (101). Hexanal (HEX) is also a product of lipid oxidation and is widely used as an index of meat flavour deterioration (92). Lipid oxidation has already been studied in food focusing on its evolution during storage (92) and more recently on the impact of thermal treatment on its formation (87, 93). However, most studies emphasize the evolution of lipid oxidation along with temperature or time, with only a few mimicking the real home-cooking methods as roasting, grilling, microwaving, among others (87, 102). Time of cooking seems to contribute more for lipid oxidation than temperature (96, 103). Roasting (200 ºC, 12 min), microwaving (1000 W, 1.5 min), grilling (130-150 ºC, 5 min), and frying (170-180 ºC, 4 min) foal steaks increased lipid

CHAPTER I. INTRODUCTION oxidation, with roasting and microwaving contributing with the highest values (TBARS) of 92.8 and 91.6%, respectively, followed by grilling and frying with 90.8 and 79.4% increase of lipid oxidation (96). The lowest lipid oxidation observed in fried samples may result from the combined reactions of products derived from oxidation processes, thus decreasing the contents of oxidation-derived products measured by thiobarbituric acid reactive species (TBARS) assay (96). Sous-vide cooking (60-80 ºC, 6-24h) of lamb loins also induced lipid oxidation (90). Moreover, all the above mentioned reports studied the oxidative reactions in beef cooked meats while chicken meat received less attention concerning the formation of lipid oxidation markers, thus, the impact of conventional home-cooking on chicken meat lipid oxidation should thus be studied.

I.2.1.3 VITAMINS AND MINERALS

Fat-soluble (A, D, E, and K) and water-soluble vitamins (B-complex and C) have different heat stability; thus, their degradation also depends on specific cooking conditions such as the cooking method, time, temperature, presence of light and oxygen, moisture content, and pH (104, 105). Vitamin content can vary in different parts of the same tissues and among animals according to the origin and time of slaughter. Although different impacts on vitamin loss are reported, most studies agree that vitamins in meat are lost during cooking; however, in the case of vitamins A and B6, an increase in vitamin content has also been described (77, 80, 81, 83). Losses of 15.4-39.3% in vitamin A were observed in grilled and boiled meat (beef and pork) (77). Microwaving chicken and lamb resulted in minor losses of vitamin A compared with the content of oven roasted chicken and lamb (106). The higher content of vitamins, found in microwaved samples, may be the result of shorter heating times required to cook the meats, resulting in less destruction of vitamins. Vitamin B1 is the most heat-labile vitamin of the B-complex vitamins. Gerber and colleagues (77) reported that 75% of vitamin B1 is lost in grilled pork, and 100% is lost in boiled beef brisket. Grilling meat also affects the contents of vitamins B2 and B3; losses of 50% in grilled and 84% in boiled meat are described for these two vitamins (77, 80). Because these are water-soluble vitamins, these losses may be the result of vitamins leaching into the drip (grilling) or water (boiling) (77, 80). Vitamin B6 is also lost during baking, microwaving, and frying, but increases after grilling (80). Vitamin B12 is relatively resistant to higher temperatures, with major losses occurring on the surface of the meat caused by the direct contact with heat during cooking (84). Roasting and grilling have little effect on the content of vitamin B12, while frying reduces the content of this vitamin by 32% compared with that in raw meat (107). No reports were found regarding the stability of vitamin C in cooked muscle tissue of meat. Little is known about the effect of cooking

CHAPTER I. INTRODUCTION methods on the stability of vitamin E in meat. Frying or grilling sausages does not affect the content of α-tocopherol (108), while grilling meat (beef, pork, and veal) reduces the content of vitamin E by 11-21.8% (77). No studies were found on the stability of vitamin D and K in cooked meat. Few reports address the effect of cooking on the mineral content of meat. Boiling pork loin increases the mineral content as the consequence of water lost during cooking. The internal temperature endpoint, during boiling, significantly affects mineral content (109). Grilling and boiling pork and beef decreases the contents of Na, K, P, Ca, and Mg, while increasing the contents of Fe and Zn (77). Deep-frying, pan-frying, oven cooking, and microwaving decrease the mineral content of cooked beef steak by 13.6-21.1%, with microwaving causing the highest loss (110). Purchas and colleagues (111) compared the mineral content in uncooked and cooked lean beef, and reported a decrease in the contents of Na and K, and an increase in Ca, Cu, Fe, Mn, and Zn, in cooked meat compared with the levels found in raw meat. These results indicate that divalent minerals are better retained during cooking than Na and K; this may be because their greater association with protein prevents the decrease in the contents of divalent minerals during cooking.

I.2.2 CONTAMINANTS STABILITY TO COOKING

I.2.2.1 FOOD PROCESSING CONTAMINANTS

Besides oxidizing nutrients, cooking meat at high temperatures, such as barbecuing, roasting, and grilling, may lead to the formation of other undesirable compounds, such as heterocyclic aromatic amines (HAAs) (112-115). Additionally, barbecuing over charcoal contributes to the formation of polycyclic aromatic hydrocarbons (PAHs) (116, 117). Processed meat also contains N-nitroso compounds because of preservation with nitrites. Epidemiological studies indicate that high consumption of well-done meat, particularly that of red meat, is related to the increased risk of colon, pancreatic, gastrointestinal, lung, liver, prostate, skin, and breast cancers, which are related to higher exposure to HAAs and PAHs (118). The IARC has recommended reducing human exposure to these foods (20). The HAAs are divided in two classes: aminoimidazole-azaarenes and amino-carboniles. The former, also known as “thermic HAAs” or IQ type amines, are formed at temperatures between 150-250 °C and result from complex reactions that involve creatine, free amino acids, and carbohydrates, while the second are called “pyrolytic HAAs” or non-IQ amines and are produced from pyrolysis of proteins or amino acids, such as tryptophan and glutamic acid, heated at high temperature (above 300 °C). Due to their high carcinogenicity, 16

CHAPTER I. INTRODUCTION minimal exposure to these compounds is recommended (114, 119). The production of HAAs in meat has been reviewed by Shabbir and colleagues (120). The extent of HAA formation in cooked muscle foods depends on the type of meat (bovine, pork, or poultry), the age of the meat (time post-mortem), cooking process (time and temperature), fat and moisture contents, acidity, HAA precursors (free amino acids and hexoses), degree of browning, and doneness (12, 112, 113, 117, 119, 121-123). The type of charcoal, used for barbecuing, also influences the formation of HAAs. Continuous use of the same charcoal for grilling increases the formation of HAAs because of fat combustion that occurs during grilling (9). The addition of oils or solid fats while frying influenced the formation of HAAs (124), and reusing oils increased the amounts of HAAs formed in fried meat. This is likely due to free radicals, generated during the lipid peroxidation of hot frying oil, which may catalyse the condensation reaction needed for the formation of HAAs (125). PAHs presence in food can stem from environmental pollution or from cooking practices and industrial food processing such as the smoking of meat. There are two types of PAHs, light and heavy, depending on the number of fused aromatic rings in its structure (126). These compounds are widely distributed in the atmosphere and, due to their lipophilic characteristics, tend to accumulate in the food chain (127). Dietary intake of PAHs is believed to be the main source of human exposure to these pollutants (128, 129). In 2005, the FAO/WHO Expert Committee on Food Additives (130) designated 14 PAHs as genotoxic and carcinogenic compounds, and a few years later, based on data on occurrence and toxicity, the EFSA selected 8 priority PAHs as the most suitable indicators of carcinogenicity (131). Several authors have reviewed the distribution of PAHs in the environment and food chain, the potential health hazards associated with these compounds (132-134), and how to reduce their formation in smoked meat (135). PAH formation is influenced by the type of meat (bovine, pork, or chicken), as well as type of grilling, such as charcoal or gas; charcoal grilling contributes to greater PAH formation resulting from incomplete combustion of charcoal, compared with that generated by gas grilling (117, 136). Wood grilling meat shows higher PAHs compared with coconut grilling. The distance between meat and the heat source, and the temperature of the heat source are major contributors to PAH formation/transference to food (9, 137). Cooking methods, involving grilling or other foods with intense heat over a direct flame, contribute to fat dripping on the hot fire, leading to the formation of PAHs and their consequent adhesion to the food surface (9, 127).

CHAPTER I. INTRODUCTION

I.2.2.2 ANTIBIOTICS

The prevention and/or detection of the presence of antibiotics in meat requires the examination of the changes in, and stability of, these residues during cooking. Heshmati (138) and Tian and colleagues (139) reviewed the effect of thermal treatment on antibacterial drug residues in food, although they did not focus on a specific type of food such as the muscle tissue of meat. The effect of cooking meat has already been studied with respect to most antibiotics including tetracyclines (TCs), sulfonamides (SAs), quinolones (QLs), macrolides, β-lactams, and nitrofurans (Table I.3). The following sections describe the impact of distinct domestic cooking methods on the above mentioned antibiotic classes.

I.2.2.2.1 TETRACYCLINES

The stability of all four TCs – tetracycline (TC), oxytetracycline (OTC), chlortetracycline (CTC), and doxycycline (DC) - after boiling, roasting, and microwaving has been investigated, using different temperatures and times, in chicken thigh and breast samples (140) (Table I.3). Increasing cooking time increases the loss of TCs, with DC being the most heat-stable, and OTC the most heat-labile. To ensure 90% destruction of TCs in chicken meat, it is necessary to cook chicken for 23.9, 53.2, and 101.6 min using microwaving (2450 MHz), boiling (100 °C), and roasting (180 °C), respectively. This means that common cooking practices may not be enough to ensure the destruction of these antibiotics. The effects of the aforementioned cooking methods on DC residues in the edible tissues of poultry were also studied by Javadi (141) that evaluated a reduction in DC residues using the technique of microbial inhibition. The exceptions were the boiled samples, in which DC was excreted from the tissue into the cooking fluid. Time and temperature are the major agents influencing the reduction of antibiotic residues during cooking. Microwaving, frying, and boiling chicken also contribute to the loss of OTC by 74, 48, and 35%, respectively. The reduction in OTC during frying may be caused by the antibiotics leaching out of chicken meat during water loss (142). Grilling and roasting are the most and least effective cooking techniques, respectively, for reducing the levels of OTC and DC in chicken meat. Beyond assessing the effects of cooking on the reduction of tetracyclines in muscle, it is important to investigate the possible formation of degradation products that can exert even more toxicity than the parent compound. The degradation of tetracyclines can generate 4-epimers (4eTC and 4eATC) and anhydrous products (ATC) depending on temperature and pH (143). Microwaving and boiling diminish the TC content while ATC and 4eATC (144) increase, thus, although cooking causes a degradation of antibiotics, it 18

CHAPTER I. INTRODUCTION contributes to the formation of degradation products. This is of concern, because exposure to ATC and 4eATC is related to the Fanconi-type syndrome, a reversible renal dysfunction (145). The formation of degradation products (α-apo-OTC and β-apo-OTC) from OTC during microwaving and boiling of chicken and pork meat was reported by Nguyen and colleagues (146), and their toxic effects were assessed on rats for 90 days. The results suggest that the toxic effect of β-apo-OTC can damage the liver and kidney tissues of rats, and cause degeneration and necrosis of hepatocytes (146). Heated OTC, assessed with the Ames test using Salmonella strains TA98 and TA100, showed mutagenic effects compared with controls that did not have effects (147). Other studies reported the formation of unidentified compounds resulting from the degradation of tetracyclines; it is unclear whether these compounds correspond to the metabolites described previously (4-epimer, anhydro-TCs, and apo-TCs) or whether they are new products of degradation (144, 147).

I.2.2.2.2 β-LACTAMS AND MACROLIDES

To the best of our knowledge, literature, from the last decade, contains only 1 study on how cooking affects β-lactam residues in the edible tissue of poultry, and 3 studies on macrolide residues in cooked meat (Table I.3). These studies were conducted using the microbial inhibition assay (148), showing that roasting (200 °C, 40 min) reduces the contents of β-lactams, tetracyclines, and macrolides to values below the MRL. However, this test only indicates whether the levels of antibiotics are, or are not, reduced sufficiently to create or not create a zone of inhibition; the real concentration of antibiotics, after roasting, remains unclear. Additionally, it is not reliable to compare the absence of the inhibition zone with MRL levels because there are no concentration values to compare. In this sense, this test does not confirm that roasting is effective for degrading β-lactams. It is important to investigate the real effect of cooking on this class of antibiotics, because, for instance, amoxicillin was detected in the tissue of chickens in Bangladesh (149). If regulations are not respected, this class of antibiotics may present a serious threat to human health because it is used in the veterinary treatment of pigs and poultry. In Europe, this class of antibiotics has a lower MRL (50 µg/kg) compared with those of other antibiotics. Recently, Hussein and colleagues (142) used a more accurate technique of high- performance liquid chromatography (HPLC) to evaluate the effect of cooking on tilmicosin residues in chicken meat. In this study, boiling, frying, and microwaving reduced the content of tilmicosin residues in chicken meat by 36.5, 46.4, and 41.0%, respectively. The stability of tylosin A during cooking procedures was also assessed in chicken meatballs, demonstrating that the percentage of reduction in the levels of tylosin A was dependent on

CHAPTER I. INTRODUCTION the initial concentration prior to cooking (150). Additionally, a relationship between the reduction in the content of tylosin A increase in internal temperature, and weight loss during cooking was observed. Despite the ability of cooking to reduce macrolide residues in meat, it is important to control the administration of veterinary drugs and allow sufficient time for withdrawal (35).

I.2.2.2.3 QUINOLONES

In the last decade, few studies addressed the impact of domestic cooking practices on quinolones stability in meat (Table I.3). Frying, boiling, grilling, microwaving, and roasting appear to exert some impact on quinolone residues in meat (151, 152). Compared with the levels in uncooked samples of thigh and breast chicken muscles, the levels of enrofloxacin in the same muscle groups were reduced by boiling and microwaving, and were increased by oven roasting and grilling due to the decreased moisture content contributing to the apparent concentration. Despite the observed reduction in the levels of enrofloxacin residues, the authors concluded that cooking procedures do not influence the concentration of enrofloxacin, and that uncooked chicken meat can be used to estimate the exposure of consumers to enrofloxacin because this antibiotic is stable during cooking. Ciprofloxacin, the main metabolite of enrofloxacin, is nearly completely reduced (by 96%) after frying and boiling of broiler chickens for 10 min (151). There appears to be no information on the degradation products of quinolones in cooked meat; however, Junza and others (2014) reported the formation of 27 thermal transformation products of 4 quinolones, and 24 metabolites of enrofloxacin, in cow milk, evidencing that cooking meat contaminated with quinolones may lead to the formation of compounds of unknown identity and toxicity.

Table I.3 - Cooking methods and their effects on the reduction of antibiotics (%) in meat. C Tetracyclines Macrolides Quinolones Sulfonamides Nitrofurans Refs HAPTER

TC OTC CTC DC TILM TYL CIP ENR SDZ SMX SMZ SQX AMOZ AOZ ADH SEM

(142, I.

48.6- 13.9- I Frying - - - 46.4 - 96.7 ------151, NTRODUCTION 99.1 66.3 152) (146, 6.25- 10.4- 11.8- 0.4- Deep-frying 53.9 44.9 34.3 61.8 ------153, 55.8 41.7 34.1 33.5 154) (140,

142, 25.8- 35.2- 18.9- 11.5- 4.4- 43.2- Boiling 36.5 96.0 22.4 5.51 9.1 8.47 - - - - 146, 75.6 99.0 73.8 54.7 35.3 62.8 150-152, 154) (140, 46.0- 61.2- 32.0- 13.7- +44.1- Roasting ------+187 +20.9 7.6 +697 152, 86.4 97.6 85.6 44.5 310 155) (140, Baking - 30-40 ------152, 155) +41.6- (152, Grilling ------+164 +16 10.7 +630 101 155) (140, 142, 146, 25.8- 51.5- 18.9- 11.5- 61- 47.2- Microwaving 41.0 - 61.7 61.9 51.8 54.8 +203 +24.6 +1 +741 150, 75.6 91.5 73.8 61.8 86.9 77.0 152, 154, 155)

The values expressed with a symbol (+) mean a percentage increase instead of decrease. Cooking conditions: Frying – 170-200 ºC; deep-frying – 170-190 ºC; Boiling –

100 ºC; Roasting – 170 - 200 ºC; Baking – 200 ºC; Microwaving – 100 – 800 W.

CHAPTER I. INTRODUCTION

I.2.2.2.4 SULFONAMIDES

Sulfonamides antibiotics are N-substituted derivatives of p-aminobenzenesulfonic acid, and are one of the oldest and most widely used classes of antimicrobial drugs for veterinary purposes. This class of antibiotics is composed of at least 40 different p- aminobenzenesulfonic acid derivatives, all used as veterinary drugs (34). Sulfadiazine (SDZ) is one of the most used sulfonamides for animal treatment, usually administered along with trimethoprim (TMP). Javadi and colleagues (156) studied the effect of roasting, boiling, and microwaving on SDZ and TMP residues, in the edible tissues of broiler chickens, using the microbial inhibition method that involves seeding plates with Bacillus subtilis at pH 7.2. All cooking procedures reduce the diameter of the inhibition zone compared with that in the raw samples; microwaving and boiling are the most and least effective methods, respectively. Sulfonamides (SAs), such as SDZ, sulfamethazine (SMZ), sulfamethoxazole (SMX), and sulfaquinoxaline (SQX), are decreased after boiling, deep- frying, and microwaving chicken meatballs (153) (Table I.3). Both time and temperature greatly contribute to the reduction of all SAs. The efficacy of all cooking methods, in reducing SA residues, is designated as deep-frying > boiling > microwaving, with SDZ being the most heat-labile sulfonamide. The higher reduction caused by boiling, compared with that caused by microwaving, may result from SAs leaching from meatballs into water during boiling because SAs are water-soluble. Additionally, the reduction in SAs positively correlates with the increase of internal temperature during deep-frying, and with the increase of time and temperature used for deep-frying. The maximum reductions in the levels of SAs are 37.5, 27.5, 40.7, and 27.6% for SDZ, SMZ, SMX, and SQX, respectively (Ismail-Fitry and others 2008). Thus, cooking time and temperature appear to be the main factors affecting the reduction in the levels of SAs in meat. However, the products of sulfonamide degradation during cooking remain unknown.

I.2.2.2.5 NITROFURANS

Despite being banned since the 1990s, nitrofurans still pose a threat to human health because of the possible continuous misuse of this antibiotic in livestock production. After administration, the nitrofurans furazolidone, furaltadone, nitrofurantoin, and nitrofurazone are rapidly metabolized to 3-amino-2-oxazolidinone (AOZ), 3-amino-5-morpholinomethyl-2- oxozolidinone (AMOZ), 1-aminohydantoin (AHD), and semicarbazide (SEM), respectively. Furazolidone residues are decreased post-mortem because the degrading enzymes that convert furazolidone into its metabolites, remain active for a long period after animal death (157). From our literature research, only two reports from the last decade were found 22

CHAPTER I. INTRODUCTION regarding the impact of meat cooking on nitrofurans. Cooper and Kennedy (155) found that frying, grilling, roasting, and microwaving does not affect nitrofuran metabolites in the muscle and liver tissues of pigs (Table I.3). Shitandi and colleagues (158) also reported furazolidone as highly stable during deep-frying, being a risk to human health. In this case, no degradation products, derived from nitrofuran metabolites are expected to be formed because of the high stability of nitrofurans during the cooking process.

I.2.2.3 MYCOTOXINS

Thermal processing has been used as a mitigation strategy to reduce mycotoxins content in cereals, even though these metabolites have been described as resistant to thermal treatments (159-162). However, they may be degraded depending on temperature, time and moisture conditions during processing (161). Literature reports higher degradations of aflatoxins (AFs), DON, T2, OTA, and fumonisins (FBs) in food products with higher moisture content, whereas ZEN stability seems not to be influenced by moisture content (159-163). Literature shows contradictory results, some studies report mycotoxins either as highly stable to high temperatures, whereas others describe them as thermo-labile (160-162). However, in general, the information concerning the impact of domestic cooking on mycotoxins stability in muscle foods is scarce. So far, only three reports were found: i) Furtado and colleagues (164) who evaluated the impact of cooking and/or processing on aflatoxin B1 and B2 stability in meat from pigs fed with a contaminated diet; ii) Pleadin and colleagues (6) who evaluated the impact of frying, boiling and baking on OTA’s stability in meat sausages; and, iii) Tolosa and colleagues (165) who reported the impact of boiling, oven cooking, and microwaving on emerging mycotoxins in fish muscle. Concerning the two studies performed on meat, the first evaluated the impact of frying pork bellies (171 ºC, 3 min) and broiling pork chops in an electric oven to an internal temperature of 76 ºC, reporting aflatoxin B1 and B2 as stable to cooking, with reductions ranging from 15 to 30% (164); and the second evaluated the impact of frying (170 ºC, 30 min) and boiling (100 ºC, 30 min) on OTAs stability in meat sausages concluding that this toxin was poorly lost after cooking with reduction rates of 12.6% and 7.4%, respectively. In the later study, only increased cooking time (~60 min) resulted in significant reductions (75.8%) (6). More research is needed regarding mycotoxins fate after cooking contaminated meat taking into account that only 3 mycotoxins were evaluated, two of them belonging to the same chemical group and only 4 cooking methods were tested.

CHAPTER I. INTRODUCTION

I.3 EXPOSURE TO DIETARY CHEMICAL HAZARDS

Food intake is considered a major route of exposure to many contaminants, in human health risk assessment. The beneficial or deleterious effects of a compound cannot be based on its content in raw or cooked meat, because the amount of a nutrient or a contaminant present in meat does not directly reflect the actual amount released (bioaccessible) from the food matrix into the gastrointestinal tract (GIT) during digestion. After ingestion, meat suffers several chemical reactions along GIT. The digestion process is very complex involving four distinct compartments – mouth, stomach, small intestine, and large intestine – with the ability to convert different types of food into nutrients at molecular scale able to be absorbed in a controlled matter (166). When food enters the oral cavity the digestion process begins: the solid foods are chewed to reduce particle size and saliva is released containing amylase and lubricating fluids to allow food swallowing and passage to stomach; then, in the stomach, the proteases – pepsin and gastric lipase – start digesting the bolus under acidic environment (pH 3) while the gastric emptying releases the chyme into the small intestine. Once the chyme reaches the duodenum, the pH rises towards neutral and pancreatic proteases, lipase, and amylase are secreted along with bile to digest the nutrients and allow their absorption. The small intestine is the GI compartment where most of the digestion and absorption takes place; it comprises the duodenum, jejunum, and ileum, and chyme passes from one to the other until reaching the large intestine where undigested parts of food are partially fermented by gut microbiota (166). The oral bioavailability of a compound is usually divided in three processes: i) the bioaccessibility that determines the release of compounds from its matrix in lumen; ii) the transport across gastrointestinal epithelium in the body; and iii) metabolism in liver.

I.3.1 BIOACCESSIBILITY

The term bioaccessibility defines the quantity or fraction of a compound that is released from the food matrix during digestion and becomes available to be absorbed (167). The bioaccessible content is always equal or higher than the bioavailable portion that is able to reach the systemic circulation, thus bioaccessibility levels usually estimate the maximal bioavailability for any nutrient or contaminant (167, 168). The bioaccessibility can be estimated using in vivo or in vitro models, with in vivo studies having major drawbacks mostly due to ethical constraints (167). In this sense, a wide range of gastrointestinal systems have been designed to study the ingestion and digestion process of nutrients and contaminants involving in vitro static or dynamic digestive models, 24

CHAPTER I. INTRODUCTION with the static models being preferred as these are cheaper, easier to use and allow the digestion of a higher number of samples simultaneously. These models aim to simulate the in vivo physiological conditions of the upper GIT (oral, gastric and duodenal phases). More than 2500 research articles have been published in the last 40 years using in vitro digestion assays, with more than 85% of the papers being published in the last two decades (169), showing the current interest of the scientific community in the in vitro digestion of foods to comprehend the risk assessment of the ingested compounds. Nowadays, in vitro digestion has become a valuable and convenient tool in studying chemical changes, digestibility and even toxic compounds under simulated gastrointestinal conditions. Although many studies (87, 93, 96, 102) have reported the production of oxidized lipids and proteins in cooked meat, as refered in previous sections, the interaction between food and oxidized components during simulated digestion remain to be elucidated. The oxidation processes are likely to be emphasized along GIT. Several authors have revealed that lipids can be oxidized under in vitro (170-172) and in vivo digestion conditions (173). The stomach has been pointed out as a bioreactor of many chemical and biochemical reactions (174), including lipid peroxidation and consequent oxidation of dietary constituents, such as proteins. Some studies revealed that meat in the stomach generates free radicals promoting further lipid and protein oxidation reactions (175, 176). MDA and HNE levels in blood and urine were highly affected by dietary factors and potentiated by peroxidation during digestion (177, 178). Depending on the degree of oxidation, these processes may impair the enzymatic activity of proteases during digestion and decrease the protein digestibility (86, 179, 180). The temperature of cooking is a determinant factor for protein digestion rate, increasing pepsin speed digestion at lower temperature (70 ºC) by unfolding of protein allowing the enzyme to access the active site, while higher temperatures slowed down pepsin activity (86). Sante-Lhoutellier and colleagues (86) and Bax and colleagues (179) showed that cooking temperatures above 100 ºC led to protein carbonylation and aggregation decreasing pepsin enzymatic activity. An ineffective enzymatic digestion of meat may compromise the overall protein quality by reducing amino acids bioavailability, principally those that are essential, as well as it could possibly contribute to the formation of toxic compounds namely MDA, HNE, and 4-hydroxy-2- hexenal (HHE) during digestion. The use of in vitro digestion methods has become a valuable and convenient tool to evaluate the digestibility and formation of toxic compounds under simulated in vitro digestion conditions (170, 171). Additionally, food matrix and pre-treatments applied (e.g. thermal treatment or seasoning) may have a significant impact on bioaccessibility of chemical hazards (3, 168), thus understanding the influence of ingredients on their final content before absorption is a

CHAPTER I. INTRODUCTION relevant matter. Eating food leads to physiological changes in the GIT, which may have the most significant impact on bioaccessibility of a compound. This was demonstrated by Versantvoort and colleagues (181), which observed an 8-fold increase of the bioaccessibility of benzo[a]pyrene from soil when simulating fed conditions instead of fasting conditions. In the case of antibiotics, several factors have been pointed out to influence their stability throughout all GIT and consequent absorption, such as solubility issues, drastic pH switching conditions between oral (pH 7)/ gastric (pH 3)/ intestinal (pH 7) phases, and bile acids (182, 183). The food matrix may also impact the final bioaccessibility of compounds, either enhancing or decreasing their contents (184). So far, no studies were found concerning the impact of culinary practices on the bioaccessibility of antibiotics in contaminated meat. Concerning the mycotoxins, their bioaccessibility in cereal matrices have been extensively studied (185-189) showing that most of these toxins are bioaccessible (54 to 108%). However, it seems that adding protein ingredients strongly decreases mycotoxins bioaccessibility in wheat crispy bread (190). So far, no information was found regarding mycotoxins bioaccessibility in meat matrices. In this sense, a wide range of matrices must be studied concerning the bioaccessibility of mycotoxins since most studies published are focused on carbohydrate-rich matrices due to the natural and predominant occurrence of these toxins in such matrices. Nevertheless, the mycotoxins are also present in other foods, such as meat products that are protein-rich matrices, so their bioaccessibility as well as the impact of cooking methods on the in vitro bioaccessibility should be explored.

I.3.1.1 IN VITRO DIGESTION MODELS

The in vitro digestion models are preferably used over in vivo studies since it reduces the use of experimental animals and facilitates the possibility to investigate a large number of samples and variables in standardized settings (189). The in vitro models have been applied in a wide range of foods with meat products and marine food comprising 25% of commonly studied food products (169). These in vitro models attempt to recreate the aspects of human GIT physiology concerning the chemical composition of digestive fluids, pH, and resistance time periods typical for each compartment. These models can include a single or multi-compartmental static or dynamic systems, and may differ in the number of steps of digestion sequence, but usually including the oral, gastric and small intestinal phases. In some cases they also contain the large intestinal fermentation (169). Lucas- Gonzalez and colleagues (169) recently reviewed the main advantages and disadvantages of dynamic and static digestion methods concerning their specific application. The ability of

CHAPTER I. INTRODUCTION dynamic models to reproduce the peristaltic movements, gastric emptying, control and adjustment of pH, temperature, and the secretion flow rates that occur during human digestion makes these models more suitable to simulate the digestion of foods for a wide range of purposes. Moreover, the mobilization of a compound from its matrix during digestion is a dynamic process with continuous change in physiological conditions in GIT, while in static models the transition of one compartment to the other is instantaneous (181). The choice of the most suitable model to use relies on the type of information required, such as, the use of a simple, cheap, and easy method to handle many samples should be used for screening purposes, while more detailed information concerning the release of a compounds from matrix are better provided using a dynamic model, although these are complex, require advanced laboratory equipment, expensive to set up and maintain, and may not be available to the majority of research groups (191). Moreover, only small sized samples can be introduced in the model and only one matrix can be analysed per day, unabling the analyses of many samples (181). These are the reasons why most research on in vitro digestion of food is made using static models. The static models of human digestion have been used to study the digestibility and bioaccessibility of macronutrients (e.g. protein, carbohydrates, and lipids), micronutrients (e.g. minerals and vitamins) as well as contaminants (e.g. mycotoxins and heavy metals) as reviewed by Minekus and colleagues (192). These are simple models using constant ratios of food to the enzyme/electrolytes, constant pH for each digestive phase, and specific time and temperature conditions of incubation (169, 192). In the gastric phase pepsin is used to hydrolyze the homogenized food at a specific pH, time and temperature (pH 1-2, 2h, 37 ºC). Then, pancreatic enzymes and bile salts are added to simulate the intestinal phase (pH 6-7, 2h, 37 ºC) (192). Despite the simplicity of these models and the non- reproduction of all dynamic aspects that occur in GIT, they demonstrated their usefulness to predict outcomes of in vivo digestion (191, 193). However, several static models have been implemented over the years making it difficult to compare results between different research groups. The differences in pH, ionic strength, mineral type, source of enzyme and concentration used, and the time of digestion may considerably influence the results. To overcome this problem, the COST INFOGEST network proposed a standardized static in vitro protocol for the simulation of food digestion (191, 192) to ease the comparison of results among research groups. The first version of the protocol was well received by the scientific community having more than 1154 citations (Web of science) since 2014 (year of publication). However, Rodrigues and colleagues (194) suggested an update in the protocol concerning the study of lipophilic compounds (e.g. carotenoids and plant sterols) in food to allow the separation of micelles containing the lipophilic compounds by centrifugation and

CHAPTER I. INTRODUCTION consequent extraction with diethyl ether. Thus, in the past year an update of the INFOGEST protocol was published (191), knowing that the standardization of in vitro static methods still requires constant updates to answer all needs concerning a wide range of compounds existing in food.

I.3.2 ABSORPTION ALONG GASTROINTESTINAL TRACT

The release of a contaminant from food after ingestion into human GIT is a prerequisite for the uptake of a contaminant in the body. Nonetheless, the fate of ingested compounds along GIT does not rely only on mimicking the digestion process to assess their bioaccessibility, but also on determining the amount of the specific compound that is assimilated across the intestinal cell wall (195). The gastrointestinal epithelium is the first target of chemical hazards upon ingestion, mostly when present in higher doses (196). Stomach and intestine are natural barriers, composed by semipermeable cells membranes, where molecules can be absorbed. A higher absorption of compounds is expected to occur at the intestine concerning the vast surface area which is responsible for 90% of compounds absorption along GIT (197). A molecule may cross the cellular barrier by transcellular transport (across the cell) or paracelullar transport (between cells, via tight junctions (TJ)), the later one being a size- restricted passive diffusion transport. On the other hand, the transcellular transport may include passive diffusion, carrier-mediated transport (facilitated or active) and endocytosis (198). The epithelium is also reinforced with efflux transporters such as P-glycoprotein (P- pg), also known as multidrug resistance protein 1 (MDR1) or ATP-binding cassette (ABC), and breast cancer resistance protein (BCRP) that send unwanted substances (e.g. xenobiotics) back to lumen (199). Several pro-carcinogens and dietary toxic compounds have been identified as substrates for ABCG2, among them HAAs, as PhIP (2-amino-1- methyl-6-phenylimidazo-[4,5-b]-pyridine), and mycotoxins, such as AFB1 and OTA (200). The transport of a substance across the intestinal epithelium by passive diffusion is mainly dependent on physico-chemical properties of the molecule like the lipophilicity, molecular weight, hydrogen bonding, and polar and nonpolar surface area, as well as the pH/pKa, and length of the pathway (201). According to Versantvoort and colleagues (181), it is not expected the matrix to have an effect on the absorption of a contaminant. Nevertheless, in some cases the matrix of ingestion has shown to affect the transport of the contaminant across the intestinal epithelium. For example, metal and minerals seem to compete for transport across the intestinal epithelium (181). Experimental studies in animal and humans suggest that the oral bioavailability of compounds may differ depending on

CHAPTER I. INTRODUCTION food source, food preparation and processing. This means that different food matrices may confer different toxic levels for the same contaminant (181). The exposure to hazardous compounds that are usually highly bioaccessible may suggest a long time exposure (4h) at gastric and intestinal levels which likely promotes the absorption of these toxins. Moreover, the long exposure period, even at low doses, may exert toxic effects along the GIT (202). Most of literature research focuses on the intestinal uptake of chemical hazards because 90% absorption is believed to occur in duodenum, ignoring stomach as the first barrier met by hazards to be absorbed. The intestinal uptake of oxidized compounds has been investigated by several authors (203-205). Maestre and colleagues (204) compared the transport of oxidized and non- oxidized ω3-PUFAs across Caco-2 monolayers, reporting that the uptake of oxidized PUFAS was around 10% of the uptake of non-oxidized PUFAS. Moreover, an increase of PUFAs uptake was observed with the addition of antioxidant polyphenols (protection against PUFAs oxidation) (204). On the other hand, the uptake of linoleic acid or its oxidized form did not differ across Caco-2 cells, according to Penumetcha and colleagues (205). Moreover, Dasilva and colleagues (206) recently showed that the combination of PUFAs also affect their stability and consequent oxidation along GIT, further influencing their final uptake. HNE and HHE also showed to be absorbed across a co-culture of Caco-2/TC7 cells with a maximum absorption of 1%, with HHE being preferably absorbed than HNE (203). Concerning the antibiotics used in animal livestock, these are poorly absorbed in gut of animals, with 30-90% of parent compounds being excreted in feces or urine (207). In humans, tetracyclines were reported as poorly absorbed in fed conditions. These drugs have a high affinity to chelate with metallic cations forming complexes either insoluble or poorly absorbable in GIT. Some foods, such as milk, have shown to interfere by 50-90% in TCs absorption (208, 209). The same way fluoroquinolones chelate with multivalent cations in GIT (210). Moreover, co-administration of milk or yogurt reduced by 50% the absorption of norfloxacin in humans (211). On the other hand, sulfonamides absorption was not reduced when administered with food to humans, only suffering a delay of absorption rate (212). Mycotoxins have been studied regarding their absorption at intestinal level, and described to rapidly reach blood circulation suggesting that their absorption takes place in the upper part of gastrointestinal tract (213). However, knowledge concerning human gastric absorption of mycotoxins is missing, being only found information on the absorption of patulin and OTA from rat stomach (214, 215) and ergot alkaloids in ruminant gastric tissues (216). According to EFSA reports some mycotoxins (AFs, OTA, DON) are rapidly absorbed, distributed and excreted after administration using animal absorption models

CHAPTER I. INTRODUCTION

(217-219), while others (FBs) are poorly absorbed, following a rapid distribution and excretion (220). Moreover, the bioavailability of some of these compounds differ among animal species (213), and knowledge concerning mycotoxins humans’ absorption is very limited. Concerning in vivo absorption evaluation in human’s data is not available for trichothecenes, FBs, OTA however the IARC monographs include relevant information concerning AFB1 in vivo absorption in humans (221-223).

I.3.2.1 IN VITRO ABSORPTION MODELS

The bioavailability quantification is difficult and often vulnerable by complex processes comprising digestion and absorption. The best information on relative bioavailability of ingested compounds should be obtained from accurate in vivo experiments in man or animals. However, these techniques require a vast number of experimental individuals, thus, in vitro techniques have been preferred (181). In vitro absorption models approaches have been employed to determine the amount of nutrient or contaminant that is assimilated across the intestinal cell wall (195). Simple filtration (224) or centrifugation of the in vitro digested food can be used as an easy protocol to access the level of a substance that is transported across the intestinal epithelium (181). These techniques are accepted for contaminants representing the worst-case scenario, giving an overestimated amount of the absorbed contaminant (167, 189). Dialysis membranes have also been used to mimic intestinal cell tissue (167). Nowadays, a wide range of in vitro experimental models with cost-effective and adequate absorption predictability are available to evaluate both gastric and intestinal permeability and transport of molecules (225). The parallel artificial membrane permeability assay (PAMPA), cell-based models, and tissues-based models are amongst the most popular in vitro models used to perform absorption studies (225). Cell-based models using commercialized cell lines are simple, cost-effective, confer a high-throughput screening and predictive ability, and do not have ethical implications. However, these methods may present some variability of results as they depend on culture medium, temperature, cell strain, passage number, and protocol used (21, 225). Since GIT is the most frequent route for food contaminants exposure, it is of paramount importance to assess the impact of food contaminants on gastrointestinal epithelium cells. Therefore, the use of in vitro cell culture systems with a monolayer of villus cells provides a valuable tool in the study of compound transport. These cells need to be viable, polarized, and fully differentiated and mimic the ones found in the stomach and intestine (198, 199, 225).

CHAPTER I. INTRODUCTION

Regarding gastric absorption, the transport studies using primary gastric epithelial cells are technically difficult to achieve, thus gastric permeability studies are widely performed using immortalized gastric cell lines, because they are easier to reproduce and maintain in culture. There are several human gastric carcinoma cell lines namely Hs746t, AGS, NCI- N87, and MKN28. The MKN28 cells have been used to perform gastric transport studies, since these are able to express TJ proteins – claudin-3, claudin-4, claudin-7, ZO-1, and occludin) to form a cohesive epithelial barrier (226, 227). The NCI-N87 cell line has also been used as a gastric epithelial barrier model for drug permeability assays. This cell line has a unique combination of properties: i) expresses adhesion proteins, as E-cadherin and ZO-1, ii) expresses and produces gastric mucin, lipases, pepsinogens, and zymogens; iii) is capable to form a tightly cohesive epithelium; and iv) is stable at long post-confluence stages (228). Moreover, Lemieux and colleagues (229) suggested this cell line as a potential model to predict gastric permeability and absorption of molecules under physiological conditions. Studies regarding drugs and toxins absorption are mostly focused on the intestinal epithelium using distinct intestinal epithelial cell lines from various animal species and human origin (230). Caco-2 monolayers are the widely used human cell model to predict permeability and intestinal absorption of molecules (231). This cell line was originally established in 1977 (232) and has been widely adopted by the pharmaceutical and food industries for screening purposes. The Caco-2 cells derive from a human colon adenocarcinoma lineage and after prolonged culture (up to 21 days) on semipermeable membranes (233), they undergo spontaneous differentiation resulting in polarization and formation of the TJ proteins between adjacent cells, mimicking the enterocytes of the small intestine barrier (234). In addition, Caco-2 cells also express multiple uptake and efflux transporters (235). Transport studies are usually performed on Transwell inserts systems. This is a common technique used to grow Caco-2 cells, which allows the transport assessment of drugs and toxins from the apical to the basolateral compartment, as well as in the opposite direction (Figure I.2). The inserts are composed of a 10 µm-thick membrane made of polyester or polycarbonate and are available with different pore sizes. The cells are seeded on the membrane that separates the apical side (lumen) from the basolateral compartment (epithelial layer). To monitor the integrity of the cell monolayer the trans-epithelial electrical resistance (TEER) across the cell monolayer is measured. An optimal monolayer integrity suggests that there are TJ between adjacent epithelial cells providing a good separation between the apical and basolateral sides (231). Thus, the established cell monolayer can be challenged from the apical site to the basolateral site with toxins as well as other

CHAPTER I. INTRODUCTION compounds (e.g food contaminants) allowing a wide range of functional parameters to be measured (236-238). Caco-2 cells have been widely used to study intestinal transport of mycotoxins like DON and their acetylated metabolites, FB1, OTA, aflatoxin M1, and ZEN (231, 239-245).

Figure I.2 – Schematic representation of a cell monolayer (e.g. Caco-2 cells) grown on a transwell system (based on Hubatsch and colleagues (231)).

However, the Caco-2 model cell presents disadvantages like the high variable expression of P-pg and low expression of metabolizing enzymes, namely the most important enzyme in the human gut, CYP3A4, which is low or absent in Caco-2 cells (199). To overcome these limitations, other cell lines have been used or designed as alternative to Caco-2 cells, such as, the Madin–Darby canine kidney (MDCK), the TC7 (Caco-2 subclone), the rat duodenal cell line 2/4/A1 (199), among others. In contrast to Caco-2 cells, the TC7 cells are a clonal population isolated from Caco-2 cells and exhibit high expression levels of typical functional markers, most importantly CYP3A4, the major phase I metabolizing enzyme in the gut wall (199). Notwithstanding, Caco-2 cells are by far the most used model for transport assay nowadays, thus enabling a higher degree of results comparison between research groups. Co-cultured Caco-2 with other cells may improve the experimental results, for example, the co-culture of Caco-2 with HT29 cell lines incorporates the influence of mucus upon transport of molecules. Mahler and colleagues (246) verified that the ratio of HT29- MTX/Caco-2 cells strongly influenced the ability to measure iron uptake, meaning that mucus may play a role on nutrients/contaminants uptake and the addition of HT29-MTX to Caco-2 cells model may give more accurate information about the intestinal uptake of compounds.

CHAPTER I. INTRODUCTION

I.3.3 TOXICITY

The majority of toxicological data is obtained by evaluating the effects of chemical contaminants when present isolated, however we are concomitantly exposed to a mixture of toxic compounds (247). A meal can be a cocktail of chemical hazards from different origins at once. From the chemical hazards previously mentioned, the cooking-induced hazards and mycotoxins are the ones with high toxic relevance. Although not directly linked to their toxic effects, the presence of antibiotics in meat has been restricted to a maximum residue content, to control their presence and avoid side effects in humans after ingestion (37, 248). These limits depend on the compounds, as for example, amoxicillin (β-lactam), trimethoprim, and narasin (cocciodiostat) have the lowest MRLs of 50 µg/kg in poultry muscle, while tylosin (macrolide), all sulfonamides, and all tetracyclines (sum of parent compounds and its 4-epimers) have been set to MRLs of 100 µg/kg in poultry muscle; also, the sum of both fluoroquinolones (FQs) (ciprofloxacin + enrofloxacin) must not exceed 100 µg/kg; and finally, nicarbarzin additive (coccidiostat), composed by an equimolar complex of 4,4’dinitrocarbanilide (DNC) and 2-hydroxy-4,6-dimethyl-pyrimidine (HDP), has the highest MRL of 4 mg DNC/kg of muscle. DNC substance is the only marker considered of concern from nicarbazin in chicken meat (249). As previously referred, the ingestion of cooked meat may contribute with the exposure to compounds of concern, namely HAAs, PAHs, and lipid oxidation products (LOPs) (MDA and HNE). HAAs and PAHs have been pointed by IARC as a risk factor for human cancer. In 1993, the IARC classified eight HAAs as possible human carcinogens (group 2B) and one as probable human carcinogen (group 2A). In 2010, among the PAHs, the IARC classified the benzo[a]pyrene as carcinogenic to humans (Group 1), four other PAHs as probably carcinogenic to humans (group 2A), and other eleven compounds as possibly carcinogenic to humans (group 2B) (250). Indeed, in 2015, the IARC released a monograph associating the consumption of red and processed meat as potentially carcinogenic as cooking these products has proved to produce these two classes of carcinogens (20). HNE and MDA, among other reactive aldehydes, are formed during cooking of meat as result of lipid peroxidation. These compounds have not been classified concerning their carcinogenicity to humans, however, several investigations shows that these compounds are able to form adducts with DNA and proteins. These reactions lead to the alteration in the original lipids functions and may cause several diseases, such as metabolic and neurodegenerative ones, and cancer (101). These compounds also showed to induce oxidative stress and trigger inflammation in the upper intestine of mice (203). MDA and HNE can also disturb the redox homeostasis, and be involved in several diseases such as atherogenesis, diabetes and even cancer (251, 252). HNE has also been related to liver 33

CHAPTER I. INTRODUCTION disease, Parkinson, Alzheimer, and Huntington diseases (253). In fact, the Belgian Superior Health Council has considered these two compounds as major concerns for humans’ health (254). In addition, to the best of our knowledge there is no data available on the combined toxic effect of these compounds, neither maximum advisable limits in foods. Concerning mycotoxins, at the moment, no MRLs have been established regarding the presence of these metabolites in meat and meat products. Mycotoxins’ maximum limits are mostly controlled in cereal-based matrices. In those, AFs and OTA have the lowest MRL ranging from 2-12 and 4-10 µg/kg for AFB1 and total AFs, respectively, and 2 to 15 µg/kg in the case of OTA depending if the products are intended for direct consumption or to be used in foodstuff; while higher MRL have been established for the other mycotoxins ranging from 500 to 1750 µg/kg to DON and 800 and 4000 µg/kg for the sum of FBs (FB1 + FB2). Different toxicological mechanisms of some mycotoxins at the cellular level are described in literature: (i) AFB1 causes DNA damage by forming DNA-adducts (255), affects RNA translation thus inhibiting protein synthesis, and induces oxidative stress (256); (ii) OTA is a structural analogue of phenylalanine, which contributes for protein synthesis inhibition (257) causing decreased cellular proliferation, apoptosis, impairment of barrier function and increasing membrane permeability (258); (iii) FB1 may be responsible for disruption of sphingolipid metabolism inducing lipid peroxidation, which may affect DNA integrity leading to DNA oxidized bases, due to its structural similarity with sphingoid bases, altering cell membrane and causing cytotoxicity (259); (iv) DON’s toxicity affects the ribosomal functions resulting in inhibition of protein and DNA synthesis causing decreased cell proliferation (260), as well as severely affecting the intestinal cells causing loss of monolayer integrity, resulting in reduction of viability and immune function (261). The IARC has performed the carcinogenic hazard assessment of some mycotoxins in humans. There, AFs have been classified as group 1 human carcinogen, whereas OTA and FBs were classified as probable human carcinogens in group 2B. Trichothecenes (DON and T2), ZEN, and patulin were not classifiable as to its carcinogenicity to humans which means that there are too limited, inadequate, or no data to classify them (group 3) (262). Despite that, several recent studies point to the high toxicity of this group of compounds throughout several other mechanisms as referred (196, 263-265). For example, trichothecenes and patulin have shown to affect the intestinal barrier integrity impairing the expression and function of TJ proteins in different ways, with DON being identified to modulate the expression, intracellular localization and function of TJ proteins, while patulin seems to directly affect the epithelial cell monolayer (266, 267). Moreover, these metabolites are a good example of the importance of combined toxicity evaluation as they are omnipresent in the environment and most importantly, their natural

CHAPTER I. INTRODUCTION co-occurrence in food is frequently reported (47, 268, 269). Despite the distinct mechanistic pathways of the aforementioned mycotoxins, once combined these may lead to additive, synergistic or antagonistic toxic effects after mixed exposure originated from different environmental sources via gastrointestinal and inhalation routes, for instance. Dietary exposure itself often results from the combination of distinct food items ingested together in meals, which increases even more the probability of multiple mycotoxins occurrence. The high probability of mixed exposure triggers the need for the assessment of toxicological effects of mixtures of compounds.

I.3.3.1 CELL-BASED MODELS TO EVALUATE TOXICOLOGICAL INTERACTIONS

The use of human cell models to study the impact of co-existing compounds interaction is advisable to obtain robust results in standardised thus comparable conditions. Caco-2 and HepG2 cells are both human-derived, widely used as human cell models to assess the effect of several compounds on intestinal and hepatic functions, respectively (270), representing the main organs where chemical hazards (e.g. xenobiotics) can exert their toxicity in humans. In the case of Caco-2, proliferating cells are commonly preferred for cytotoxic and cells viability assays, rather than differentiated ones, as these share some characteristics with crypt enterocytes (271, 272). Both cell models were already used to evaluate the toxic effects of mycotoxins when isolated and a few in combination. Still data on cytotoxic effects of combined mycotoxins are limited and mostly focused on a specific mycotoxin group (265, 271-274); thus the human health risk associated to multiple types of mycotoxins exposure, more probable when exposure co-occurs from different environmental sources, is still not completely understood. The experimental designs used to assess toxicological interactions of chemical hazards has been reviewed by Alassane-Kpembi and colleagues (275). The interaction of compounds is usually evaluated when their combined exposure alters the biological response in comparison with their individual effect conferring synergistic (i.e. interaction leading to a greater effect than expected) or antagonistic (i.e. interaction leading to lesser effect than expected) effects. The combination of drugs has been used since earliest days to treat diseases by reducing the doses and the toxicity of the drug through combining with another to a synergistic therapeutic effect (276). A considerable number of studies performed toxicological evaluations of mixtures using the arithmetic definition of additivity method, as reviewed by Alassane-Kpembi and colleagues (275). However, the definition of an additive effect cannot simply be the result of an arithmetic sum of two or more effects, because doubling the dose does not

CHAPTER I. INTRODUCTION automatically result in doubling the effect. For example, if a toxin A and a toxin B individually both have 60% inhibition effect, once combined their additive effect could not be 120% (277). Thus, more robust methodological approaches should be tested to determine the type of the interaction and quantify its magnitude. The theoretical biology models-based definitions of additivity are proposed as more accurate approaches, among them the theoretical biology-based combination index-isobologram method proposed by Chou (277). Joseph Chou and Paul Talatay built the so called Chou-Talatay theorem for drug combination that is based on the median-effect equation, derived from the mass law principle: 푓푎 퐷 푚 = ( ) 푓푢 퐷푚

where D is the dose, Dm is the dose required for 50% inhibition, fa is the fraction affected by the dose D, m is the coefficient of the sigmoidicity of the dose effect, and fu = 1 - fa. This equation has the unified general theory for the Michaelis-Menten, Hill, Henderson- Hasselbach, and Scatchard equations in biochemical and biophysics. This model is independent of the mode of action of the compound and just considers the potency (IC50) and shape of dose-effect curve of each compound and their mixture. With this theorem these researchers built a computer software for dose-effect analysis called CompuSyn as well as the scientific term “Combination Index” (CI) that quantifies the combined effects as additive effect (C=1), synergism (C<1), or antagonism (CI>1) in drug combinations. Furthermore, this theory also provides CI-plots and isobolograms giving an automated computer simulation for synergism and/or antagonism at any effect and dose level. This theorem has already been applied to study the interaction of two or more mycotoxins (265, 273, 274), but mostly focused on a specific mycotoxin group, as well as a combined effects of two distinct hazards, DON and cadmium, on Caco-2 cells (278). Thus the human health risk associated with multiple types of mycotoxins exposure or other chemical hazards, more probable to co-occur in different environmental sources, is still not completely understood and worth study.

CHAPTER I. INTRODUCTION

I.4 MITIGATION STRATEGIES TO REDUCE EXPOSURE TO DIETARY CHEMICAL

HAZARDS

Depending on the type of hazard compounds several mitigation strategies can be applied. Raw meat has an inherent level of contamination (e.g. antibiotics and mycotoxins) and cooking processes, although helping to reduce some initial contamination, also promote the oxidation of nutrients (i.e. formation of reactive aldehydes, carbonyls, HAAs, and PAHs). Moreover, the gastrointestinal tract may also trigger the oxidation of hazardous compounds and/or promote the bioaccessibility of other inherent contaminants that were previously trapped in the matrix. Thus strategies to decontaminate meat from food contaminants and/or reduce/avoid oxidation processes shall be employed to ensure safety and keep the nutritional value of meat.

I.4.1 COOKING-INDUCED HAZARDS

The selection of type of cooking (11, 279) and the use of other meat preparations such as wrapping (banana leafs and aluminium foil) (126) or the use of roasting bags (14) severely influences the final amount of hazardous compounds in cooked meat. Moreover, the addition of antioxidant compounds has been used to control oxidation processes generated in meat during cooking. The prevention of ROS formation during storage and cooking of meat products have been done using synthetic antioxidants as butylated hydroxyanisole (BHA), butylated hydroxytoluene (BHT), propyl gallate (PG), and tertiary butylhydroquinone (TBHQ), however these have become a concern due to their potential toxic effects. Therefore, the addition of natural antioxidants derived from plants as spices and herbs, which are rich in radical-scavenging polyphenols, have been used to replace the synthetic ones (13), conferring several advantageous since these are cheaper, more efficient, and improve the organoleptic attributes of meat products (280). Improved oxidative stability during cooking was observed in several types of meat – chicken, turkey, pork, goal and beef – with the addition of antioxidant ingredients – oregano and honey, pomegranate seed pulps, and extracts of several fruits and vegetables, such as green tea, grape seeds, Echinacea angustifolia (281-286). Moreover, the inclusion of spices and herbs in high-fat meats has shown to effectively reduce the formation of MDA, HNE, and HEX during thermal treatment (172), however, little is known regarding their influence on protein oxidation. The formation of HAAs and PAHs is also reduced with the addition of antioxidant ingredients (287-289).

CHAPTER I. INTRODUCTION

Regarding in the in vitro digestion process, MDA, HNE, and HEX formation have been evaluated concerning the addition of different types of herbs (basil, oregano, rosemary, and thyme) and spices (black pepper, cayenne pepper, cumin, curcuma, garlic, and sweet paprika) before or after cooking beef (70 ºC, 70 min). In general, these ingredients had a more limiting oxidation during digestion when added before cooking than after heating. The herbs and curcuma were more effective ingredients on reducing oxidation processes, while black and cayenne pepper and cumin exerted a moderate effect against oxidation. Adding garlic and sweet paprika before cooking had a pro-oxidant effect, with garlic keeping the same behaviour towards in vitro digestion, while sweet paprika had an antioxidant effect (172). Moreover, the potency of herbs/spices to limit oxidation processes during cooking and in vitro digestion strongly correlated with phenolic content. Depending on type of meat, i.e. low- or high-fat meat, the choice of the antioxidant compound to add in meat is determinant to obtain the desired antioxidant behaviours in meat, because Van Hecke and colleagues (290) observed that the phenolic acids (gallic, ferulic, chlorogenic, and caffeic acids) had a pro-oxidant and antioxidant behaviour at lower or higher doses, respectively, during digestion of high-fat meats; whereas ascorbic acid behaved as a pro-oxidant at all doses. On the other hand the lipophilic compounds (α-tocopherol, quercetin, and silibinin) exerted a clear antioxidant effect while digesting high-fat beef. Additionally, adding spice mixtures (composed by cloves, cinnamon, oregano, rosemary, ginger, black pepper, paprika, and garlic) before cooking beef hamburgers reduced MDA contents in meat, plasma, and urine of men after ingestion (291, 292). Moreover, the authors also proposed that cooking a hamburger with a polyphenol-rich spice mixture may lead to potential cardiovascular benefits in patients with Type 2 diabetes mellitus since the presence of spices improved the postprandial endothelial dysfunction in men with Type 2 diabetes (291). Several types of marinades (wine or beer-based) have been studied with respect to their effect on reducing the formation of hazardous compounds in meat during cooking, namely HAAs and PAHs (12). Both red wine and beer marinades reduced HAAs in fried chicken (293) and beef (294, 295). PAH formation was reduced by 70% in grilled beef in pre-treated meat with acidic marinade (1.2% lemon juice). However, only one report was found concerning the use of red wine to reduce lipid oxidation products formation during cooking and in vitro digestion reporting low MDA plasma levels in healthy volunteers after ingestion of cooked turkey meat previously soaked in red wine (296). No studies were found concerning the influence of beer marinades on the formation of MDA, HNE, HEX, carbonyls and SB structures after cooking or in vitro digestion, only that incorporating melanoidins from dark beer in grilled turkey meat before digestion inhibited TBARS formation up to 62.8% when simulating the gastric phase (297).

CHAPTER I. INTRODUCTION

In this sense, the use of natural ingredients like herbs, spices, and/or marinades are encouraged to avoid further oxidation of nutrients during cooking, while improving the organoleptic properties of meat.

I.4.2 ANTIBIOTICS

The presence of food contaminants like antibiotics, whose presence results from the human administration to animals, should not represent a threat to humans’ health if the maximum allowable limits were respected or if they would not have the tendency to accumulate in the environment. However, not always the withdrawal time of antibiotics is respected before slaughter resulting in carry-over to humans of antibiotics residues from meat. At the moment, physico-chemical techniques based on advanced oxidation processes have been used to decontaminate wastewaters with antibiotics (298). Ma and Zhai (33) believe that future research on antibiotics will focus on novel/improved decontamination methods. As reviewed in Section I.2.2.2, the use of thermal processes, i.e. cooking of meat, could be used as a mitigation strategy to reduce antibiotics content in meat before ingestion, however, further investigation should be done considering the scarcely of information available. Moreover, no information was found regarding the behaviour of these hazards towards the presence of other culinary ingredients such as herbs or beer/wine marinades. Knowing that these ingredients are involved in oxidative reactions their presence could influence the stability of some antibiotics by preventing or enhancing their degradation. Moreover, knowing that food-drug interactions highly influence the availability of the antibiotics (299, 300), it is of paramount importance to understand the real influence culinary practices (i.e. different types of cooking and addition of ingredients) on the bioaccessibility of these compounds.

I.4.3 MYCOTOXINS

Mycotoxins mitigation strategies to decontaminate foods are usually applied during handling and/or storage before reaching the animals or consumers, and are commonly divided in three categories: biological, physical and chemical methods. In case of meat contamination with mycotoxins, the biological methods are applied during processing of curing or ripening of meat products, by inoculating atoxigenic strains that contribute to the maturation of the product and influence the growth of toxigenic strains by competing for space. However the high costs of these strategies limit their application (301). The thermal 39

CHAPTER I. INTRODUCTION processing, radiation, high pressure, or atmospheric control are amongst the physical methods applied to decontaminate cured or ripened meat or avoid formation of new mycotoxins (301). As previously mentioned in Section I.2.2.3, mycotoxins seem to be quite stable to thermal treatment, unless in high humidity conditions, thus the changing of curing or ripening conditions could prevent fungal growth and mycotoxin product, but this would probably influence/change the sensory quality of the final product. Then, chemical additives such as sorbic acid and proprionic acid are added to meat to control the growth of moulds and consequent mycotoxin production, however, their efficiency is limited (301). Thus, the increasing demand for food free of additives has triggered the use of natural ingredients such as essential oils or plant extracts to degrade mycotoxins. Lately, the removal of mycotoxins by botanicals is being considered and usually preferred over other chemical treatments in other foods (161). These are natural phytochemicals, biologically safe, environmentally friend, and cost-effective compounds, making them an alternative to other chemicals industrially used to degrade mycotoxins in food and feed. Aqueous extracts of plants were suggested as a safe additive to decontaminate foods and inhibit aﬂatoxins production (302-306). However, at the moment, no studies were found concerning the impact of natural ingredients on mitigating mycotoxins presence in meat, neither their combined effect with thermal processing. Literature also lacks information mycotoxins fate during digestion in the presence of these culinary practices - cooking and natural ingredients.

CHAPTER I. INTRODUCTION

I.5 FINAL REMARKS

Domestic cooking influences both nutritional and undesirable components of meat. Protein and lipids are oxidized during cooking, leading to the formation of hazardous compounds, but not all cooking methods have been tested concerning their contribution to oxidative processes. With respect to the undesirable compounds, such as antibiotic residues and cooking-induced contaminants, cooking time and temperature are the determining factors influencing the reduction (in the case of residues) and formation (in the case of cooking-induced contaminants). Most antibiotics are reduced after cooking. Tetracyclines are the compounds most degraded by frying, roasting, and boiling. However, their reduction may result in the formation of degradation products, from which little is known about. No studies were found on the thermal degradation of macrolides and β-lactams, while cooking meat containing quinolones may lead to the formation of toxic compounds. Studies concerning mycotoxins stability to cooking in meat matrices are scarce. Table I.4 shows a summary of impact of domestic cooking on nutrients and contaminants. The digestive process has proved to potentiate the formation of hazardous compounds such as HNE and MDA by triggering oxidative reactions mostly under gastric acidic conditions that may be bioaccessible for absorption into bloodstream. However, no information was found concerning antibiotics or mycotoxins fate during in vitro digestion of contaminated meat, which is highly relevant concerning their occurrence in meat and their toxic effect in intestine (mycotoxins) or their modulating impact on gut microbiota (antibiotics). Moreover, new data on absorption and toxicological interactions of several molecules is of paramount importance considering that their combined presence may increase toxicity to cells through synergistic effects. Several strategies can be used to reduce the formation of cooking-induced contaminants, namely HAAs and PAHs and LOPs such as the use of herbs or antioxidant- rich marinades prior to cooking. In the case of LOPs the influence of beer marinades has not been investigated yet. Moreover, the use of the abovementioned strategies has not been studied yet on anthropogenic (antibiotics) and environmental (mycotoxins) chemical hazards stability.

CHAPTER I. INTRODUCTION

CHAPTER II – MATERIAL AND METHODS

II.1 REAGENTS

II.2 SAMPLES PREPARATION

II.2.1 COOKING OF SAMPLES

II.2.2 IN VITRO DIGESTION

II.2.3 CELL-BASED METHODOLOGIES

II.3 CHEMICAL ANALYSES

II.3.1 NUTRITIONAL ANALYSIS

II.3.2 COOKING-INDUCED HAZARDS

II.3.3 ANTIBIOTICS AND COCCIDIOSTATS DRUGS

II.3.4 MYCOTOXINS

II.3.5 QUALITY CONTROL/QUALITY ASSURANCE

II.4. EXPOSURE ESTIMATION

II.4.1 BIOACCESSIBILITY ESTIMATION

II.4.2 APPARENT PERMEABILITY AND FRACTION ABSORBED ESTIMATION

II.4.3 TOXICOLOGICAL EVALUATION

II.5. STATISTICAL ANALYSES

CHAPTER II. MATERIAL AND METHODS

II.1 REAGENTS, MATERIALS AND STANDARDS

Reagents and materials used in this thesis were purchased from diverse suppliers, and are listed in Appendix I in Table AI.1. In the same way all the standards used are indicated in Table AI.2. The ultrapure water (18.2 mΩ cm-1) was purified by a Milli-Q gradient system from Millipore (Milford, MA, USA).

II.2 SAMPLES PREPARATION

II.2.1 COOKING OF MEAT

The experimental work during this PhD thesis was performed in two distinct research laboratories at Porto and Almeria and comprises a wide range of chemical hazards - 5 protein/lipid oxidation markers, 10 mycotoxins and 12 antibacterial and 2 coccidiostats drugs – thus, the experimental work was designed individually for each hazard group. Moreover, different chicken muscle (legs or breast) were used in the diverse studies. This means that the sample preparation and the cooking conditions were adapted for each studied hazard, as described in the following sub-sections.

II.2.1.1 COOKING-INDUCED HAZARDS

The chicken meat legs, the dried oregano herbs and the pilsner beer used for seasoning were obtained from a retail market in Porto, Portugal. The chicken legs were firstly deboned, and the muscles were minced in a domestic blender and separated into three batches: 1) control; 2) mixed with 0.2% (w/w) of oregano; and, 3) mixed with 12.5% (v/w) of pilsner beer. All three batches were left overnight (~14h) at 4 ºC. The day after, individual burgers of homogenized meat (50 g each, in triplicate) were weighted, moulded, and cooked, keeping raw samples (as is, with oregano, and with beer) as references. Two types of cooking were tested: oven cooking, with temperature set at 200 ºC (convective air, Teka, HI-435 model, Germany), and microwave at 760 W of power (Whirlpool, GT285/WH model, USA). The cooking times were previously optimized based on the time required to achieve an internal temperature of 85 ºC (microbiologically safety criteria) using a penetration thermometer (Lacor thermometer) immediately after removing from the ovens. All these conditions were cooked in triplicates (n=3). After cooking, surface liquids (moisture and grease) were slightly drained and all burgers were cooled to room temperature. The final weight was registered to calculate the cooking yield (CY = 100 x cooked weight / raw weight) 45

CHAPTER II. MATERIAL AND METHODS

as described by Oliveira and colleagues (307). Then, the burgers were homogenized (10,000 rpm, 10 s, Grindomix GM 200, Germany), and aliquots for each analysis were taken and stored at -20 ºC, until analyses.

II.2.1.2 ANTIBACTERIAL AND COCCIDIOSTATS DRUGS

The chicken breast muscles, dried oregano herbs and the Pilsner beer were obtained from a supermarket in Almeria, Spain. The cooking of samples was made at two different initial concentrations of antibacterial and coccidiostats drugs (ACDs): 0.1 mg/kg of each compound, representing a realistic situation (MRL in food) or at 1 mg/kg of each compound (except NAR that was fortified at 0.5 mg/kg) to identify possible transformation products (TPs) derived from cooking. The evaluation of cooking impact on ACDs stability at the two concentrations (0.1 mg/kg or 1 mg/kg) was made in different days, sharing a similar experimental design: at the previous day before cooking, the chicken breasts were homogenized and divided in two batches: the Batch 1 used as control (non spiked samples) and the Batch 2 that was fortified at the selected concentration. To guarantee a homogenous fortification, the meat was dispersed in a tray to allow a maximum surface area and the required amount of ACDs was added to a final concentration of 0.1 mg/kg or 1mg/kg. The meat was manually homogenized, left 1h in the fridge, and then homogenized again. Raw meat samples from 5 different points (4 edges and center) were taken to analyse and verify the homogeneity. Then in both Batch 1 and Batch 2, 4 equal amounts were separated to obtain the four matrices in study: i) chicken breast; ii) chicken breast with oregano (0.5% w/w); iii) chicken breast with pilsner beer (12.5% (v/w)); iv) and chicken breast with oregano and beer (using the same amounts of oregano and beer), and left overnight at 4 ºC. The next day, the homogenized chicken breasts (spiked or non-spiked) were used to make burgers of 40 g. The cooking conditions were optimized to achieve an internal temperature of 85 ºC in the burgers. The conditions used were 200 ºC over 15 min (Binder Inc, Germany) and 600 W over 1.15 min (Carrefour Home, France) for oven cooking (Ov) and microwaving (Mw), respectively. For each cooking process four conditions were tested i) cooked (Mw or Ov); ii) cooked with herbs (MwH or OvH); iii) cooked with beer (MwB or OvB); and iv) cooked with herbs and beer (MwHB or OvHB). Each cooking condition was performed in triplicate.

CHAPTER II. MATERIAL AND METHODS

II.2.1.3 MYCOTOXINS

The chicken breast muscles, the roasting bags, and the commercial mixture of herbs (composed by oregano (6.8%), coriander, lemon pulp (3.5%), marjoram (3%), garlic (2.5%), rosemary, parsley, basil and thyme) were obtained from a retail market in Porto, Portugal. The chicken breasts were kept at 4 ºC until being prepared for cooking. The experiments were performed at the laboratorial scale, using 5 g of meat for each assay. To adequately simulate the domestic cooking conditions using the roasting bags, the size of the bags (38 cm x 30 cm) were reduced to 14.1 cm x 8.5 cm accordingly to the proportion meat/bag recommended by manufacturers, and sealed using a bag sealer machine (Impulse sealer PFS – 200-300 W). This downsizing of the bag required that different conditions (time and temperature/potency) were tested to achieve the same cooking point obtained after cooking a control sample in the conditions suggested by manufacturers (200 ºC over 50 min for oven, and 700 W over 15 min for microwave cooking). For that a texture profile analysis (TPA) was performed to confirm the laboratorial conditions that correctly simulated domestic cooking using the original size bag (TPA is described in Appendix I). Multivariate statistical analyses were used to compare the TPA results of control and samples cooked in small bags, pointing out the conditions to be used in 5 g samples to better simulate the texture of control (Appendix I, Figure AI.1). In the case of herbs, a proportional amount of 100 mg of herbs/5 g of chicken breast was used. Before cooking, 50 µL of a 10 mg/L mix solution of DON, T2, AFB1, AFB2, AFG1, AFG2, FB1, FB2, OTA and ZEN was used to spike the samples at 5 different places (at each edge and in center). The spiked samples were left overnight (~14h) at 4 ºC before cooking. The optimized cooking conditions used were 200 ºC over 5 min and 350 W over 0.45 s for oven cooking (Ov) and microwaving (Mw), respectively. For each cooking process four conditions were tested i) cooked (Mw or Ov); ii) cooked in the bag (MwBg or OvBg); iii) cooked with herbs (MwH or OwH); and iv) cooked with bag and herbs (MwHBg or OvHBg). Each cooking condition was performed in triplicate.

II.2.2 IN VITRO DIGESTION

The INFOGEST protocols (191, 192) offer two in vitro digestion procedures: one using individual enzymes and the other using pancreatin extract based on trypsin activity. Two different procedures were used because some research work of this thesis had already been done when the second protocol INFOGEST 2.0 was published. The study evaluating the influence of cooking on nutrients oxidation (cooking-induced hazards study) used

CHAPTER II. MATERIAL AND METHODS

individual enzymes, while the study of antibacterial and cocciodiostats drugs, and mycotoxins was performed using the pancreatin extract. The INFOGEST inter-laboratorial experiment showed that both protocols efficiently digest protein and any method could be used to perform digestion experiments (191, 308). As in vitro digestion procedures are laborious and expensive, it was decided to use pancreatin extract based on trypsin activity instead of the individual enzymes (trypsin and chymotrypsin) in the antibiotics and mycotoxins experiments, which were the ones with the highest number of samples, and use of individual enzymes when nutrients evaluation was implied as a way to better control specific enzymatic activities.

II.2.2.1 COOKING-INDUCED HAZARDS

The in vitro digestion procedure was performed according to INFOGEST 2.0 protocol (191) simulating the three phases of digestion: oral, gastric, and duodenal. The in vitro digestion experiment was only performed for oven cooked samples since no raw meat is expected to be eaten. Shortly, 5 g of sample were mixed with 4 mL of simulated salivary fluid (SSF), 25 µL of 0.3 M CaCl2, and 0.975 mL of water. The addition of salivary amylase was omitted because the matrix used does not contain starch (191). The oral bolus was incubated 2 min in a water bath at 37 ºC, under gentle agitation in a water bath with integrated horizontal shaker at 30 cycles per min. Then, 8 mL of simulated gastric fluid

(SGF), 5 µL of 0.3 M CaCl2, 0.5 mL of rabbit gastric lipase extract (containing 2400 U/mL gastric lipase and 80000 U/mL of pepsin), 0.5 mL of pepsin (3200 U/mL), 645 µL of water and 250 µL of HCl 6N to adjust pH to 3.0 were added to obtain the gastric phase. The gastric mixture was then incubated 2 h at 37 ºC, under gentle agitation. A pH adjustment to 3.0 was made after 1h of incubation. After the gastric phase, 8.5 mL of simulated intestinal fluid (SIF), 1 mL of trypsin (4000 U/mL), 1 mL of chymotrypsin (1000 U/mL), 1 mL of pancreatic lipase (80000 U /mL), 1 mL of pancreatic amylase (8000 U/mL), 2.5 mL of bile extract (160 mM), 40 µL of 0.3 M CaCl2, 3.40 mL of water and 300 µL of 4N NaOH (to adjust pH to 7.0) were added to obtain the intestinal phase. The intestinal mixture was then incubated 2h at 37 ºC, under gentle agitation. A pH adjustment to 7.0 was made after 1h of incubation. The preparation of SSF, SGF and SIF was done according to Brodkorb and colleagues (191). All digestions were performed in quadruplicate (3 replicates + blank for pH adjustment). Blanks of digestion (matrix replaced by water) were performed in duplicate. After digestion, the samples were immersed on ice for 10 min, centrifuged (5000 g, 10 min, 4 ºC) to obtain both bioaccessible and non-bioaccessible fractions, and then the protease inhibitor (1mM, final concentration, Pefabloc®SC) was added to the bioaccessible fractions.

CHAPTER II. MATERIAL AND METHODS

II.2.2.2 ANTIBIOTICS AND COCCIDIOSTATS DRUGS, AND MYCOTOXINS

The in vitro digestion procedure was performed according to INFOGEST standardized method described by Brodkorb and colleagues (191) and Minekus and Colleagues (192), adapted to 1 g of sample. Briefly, 1 g of cooked sample was mixed with 0.7 mL of SSF, 5

μL of 0.3 M CaCl2 solution and the amount of water necessary (400 μL) to achieve a swallowable bolus, simulating the oral phase, and incubated in an oven during 2 min at 37 ºC under agitation using an overhead shaker (30 rpm, Reax 2 Rotary Shaker, Heidolph, Schwabach, Germany). The addition of salivary amylase was omitted because the matrix used does not contain starch (191). Then, the oral phase mixture was mixed with 1.5 mL of

SGF, 0.32 mL pepsin (25,000 U/mL), 1 μL 0.3 M CaCl2, and 40 μL of 6 N HCl to samples, respectively, to adjust to pH 3 simulating the gastric phase. In the case of mycotoxins higher amounts of HCl were added: 350 μL (in the case of samples with herbs) or 450 μL (in the case of samples without herbs). The gastric mixture was then incubated at 37 °C under agitation (30 rpm) over 2h. After incubation, the intestinal phase was simulated (pH 7.0) by adding SIF (2.2 mL), pancreatin solution (800 U/mL based on trypsin activity) (1 mL), bile solution (160 mM final concentration) (0.5 mL), 0.3 M CaCl2 (8 μL), 4 N NaOH (55 μL), water (0.112 mL) to the gastric chime. The intestinal mixture was incubated at 37 ºC, under agitation (30 rpm) over 2 h. The preparation of SSF, SGF and SIF was done according to Brodkorb and colleagues (191). All digestions were run in triplicate plus a fourth blank tube for pH adjustment. After digestion the samples were cooled by immersion in an ice bath for 10 min and centrifuged at 5000 g for 10 min at 4 ºC to separate the soluble bioaccessible fraction from the residual fraction. In the case of antibiotics, both bioaccessible and non- bioaccessible fractions were analysed on the same day of in vitro digestion. For mycotoxins experiment, both fractions were stored at -20 ºC until further analysis. Sub-samples were taken to perform mycotoxins quantification at 2 different stages, gastric and duodenal, by collecting 200 µL of samples and replenishing the volume with the correspondent simulated fluid. Dilutions were taken into account for final calculations.

II.2.3 CELLS-BASED METHODOLOGIES

Studies regarding mycotoxins transport alone or in mixture across gastric and intestinal epithelium were performed using NCI-N87 and Caco-2 monolayers, respectively. Cytotoxic effects were evaluated towards intestinal (Caco-2) and hepatic (HepG2) cells alone or in combination. For those purposes, four mycotoxins representative of each mycotoxin group were selected: AFB1 from aflatoxins, DON from trichothecenes, OTA from ochratoxins, and FB1 from fumonisins. 49

CHAPTER II. MATERIAL AND METHODS

II.2.3.1 CELL CULTURE

Human-derived gastric cells were purchased from ATCC/LGC Standards (Barcelona, Spain). NCI-N87 cells were grown in 75-cm2 culture flasks in complete medium (CM) constituted by RPMI with 10% heat inactivate FBS, and 1% penicillin/streptomycin and incubated at 37 ºC with 5% CO2. When the cells were 80% confluent, these were trypsinized and seeded in 96-well plates (TPP, Trasadingen, Switzerland) to perform the cytotoxic assays or seeded on transwell inserts for differentiation. Human-derived intestinal and hepatic cells (Caco-2 and HepG2) were provided by the “molecular physical-chemistry” research group of the University of Coimbra and by the toxicology laboratory of the Faculty of Pharmacy of University of Porto, respectively. Caco- 2 cells were grown in 75-cm2 culture flasks in CM constituted by DMEM with 10% heat inactivate FBS, 1% NEAA, 1% glutaMAX and 1% penicillin/streptomycin and incubated at

37 ºC with 5% CO2. When the cells were approximately 80% confluent, these were trypsinized and seeded in 96-well plates (TPP, Trasadingen, Switzerland) to perform the cytotoxic assays. In the case of the HepG2 cells, these were grown in 75-cm2 culture flasks in CM constituted by DMEM with 10% FBS, 1% glutaMAX and 1% streptomycin and incubated at 37 ºC with 5% CO2. Cytotoxicity and permeability assays were performed under passages 6-23 and 41-50 for NCI-N87 and Caco-2 cells, respectively.

II.2.3.2 CYTOTOXIC ASSAYS

II.2.3.2.1 MYCOTOXINS CYTOTOXIC ASSAY AS MEASURE OF CONTROL FOR TRANSPORT ASSAY

The cytotoxicity of AFB1, DON, FB1 and OTA and their mixture on NCI-N87 and Caco- 2 cells was evaluated after 3h exposure, using the MTT (3-(4,5-dimethylthiazol-2-yl)-2,5- diphenyltetrazolium bromide (Sigma Aldrich, St Louis, MO, USA)) assay. Proliferating NCI- N87 cells (10000 cells/ well) and Caco-2 cells (10000 cells/ well) were seeded in 96-well plates (TPP; Trasadingen, Switzerland), allowing 24h for cell adherence at 37 ºC, 5% CO2, and then exposed to mycotoxins at the concentrations used for transport assay: 1.69 µM (DON), 1.60 µM (AFB1), 0.69 µM (FB1), and 1.24 µM (OTA) over 3 hours. This time of exposure was the time used for the transport experiment. Cells treated with CM alone and with maximum methanol content (0.1%) present in all samples with mycotoxins were used as controls to ensure viability of cells during the experiments.

CHAPTER II. MATERIAL AND METHODS

II.2.3.2.2 ISOLATED AND COMBINED CYTOTOXICITY ASSAYS

Stock solutions of mycotoxins (1 mg/mL) were prepared in methanol originating the following molar concentrations: 3202 µM of AFB1, 3375 µM of DON, 1385 µM of FB1, 2476 µM of OTA, and stored at 4 °C. Before each experiment, working solutions ranging from 0.625 µM to 20 µM were prepared in CM for isolated cytotoxicity assays and the concentration of mixtures used for combination experiment assays are present in Table II.1. The MTT test was performed to evaluate the isolated and combined effect of AFB1, DON, FB1 and OTA. Proliferating Caco-2 cells (1250 cells/ well) and HepG2 cells (1000 cells/ well) were seeded in 96-well plates, allowing 24h for cell adherence at 37 ºC, 5% CO2, and then exposed to mycotoxins for 72 h, alone or in binary combination. Cells treated with CM alone and with maximum methanol content (1.5%) on mycotoxins CM solution were used as controls to ensure viability of cells during the experiments. The evaluation of the isolated toxic effect of each mycotoxin on Caco-2 and HepG2 cells viability was performed by exposing the cells to 6 different levels of concentration of each mycotoxin isolated: 0.625, 1.25, 2.5, 5, 10 and 20 µM. In the case of OTA, a higher concentration was used (25 µM) to evaluate the toxic effect, only in the case of HepG2 cells, to allow obtaining a dose-effect curve for this toxin. Results from single isolated mycotoxin experiments were used to fit a dose-response curve for each compound, and to calculate the IC50 for each the Caco-2 and HepG2 cells. The cytotoxicity of combined mycotoxins was studied in binary combination: AFB1-DON, AFB1-OTA and DON-OTA by exposing the cells to 5 levels of concentration according to constant ratio combination design proposed by

Chou-Talatay method carried out at an equipotency ratio (i.e. (IC50)1/(IC50)2 ratio) so that the contributions of effects of each drug to the combination would be roughly equal (277). In the case of AFB1-DON combination in HepG2 cells, 3 additional levels of lower concentration were used, due to the highly toxic effect of this combination at medium/higher levels, making it difficult to have a dose-effect curve for the combination, and therefore being able to accurately assess the combined effect of AFB1 and DON on these cells. All measurements were taken in quadruplicate in at least two independent experiments and data were reported as mean ± standard deviation in cell viability graphs.

CHAPTER II. MATERIAL AND METHODS

DON DON

IC50 IC50

- - -

3.08 3.00 3.08 3.03 3.00 31. 31.8 31.4 31.4 31.4

OTA/ OTA/

IC50 IC50

2 (A) and HepG2 (B) cells. HepG2 and (B) 2(A)

- - -

6.50 5.00 3.25 1.65 1.00 50.0 35.0 22.0 11.0 5.50

OTA

DON

- - -

20.0 15.0 10.0 5.00 3.00 1.60 1.10 0.70 0.35 0.18

OTA

DON

OTA

AFB1

IC50

- - -

2.00 2.00 2.00 2.00 2.00 14.3 14.0 14.0 14.0 14.0

AFB1/ OTA/

IC50 IC50

- - -

12.5 11.0 10.0 5.00 2.50 50.0 35.0 21.0 10.5 5.25

OTA OTA

- - -

25.0 22.0 20.0 10.0 5.00 3.50 2.50 1.50 0.75 0.38

AFB1 AFB1

DON DON

IC50 IC50

7.09 6.99 7.04 7.01 7.04 2.23 2.23 2.23 2.23 2.23 2.23 2.23 2.23

AFB1/ AFB1/

IC50 IC50

3.10 2.86 2.43 2.14 1.42 2.70 1.35 0.68 0.34 0.17 0.08 0.04 0.02

DON DON

2 2 cells

Concentrations of the mixtures (µM) and concentrations ratios used in the combination experiment assays on Caco assayson and experiment concentrations incombination theused the (µM) ratios ofmixtures Concentrations

–

Caco HepG2cells

II.

22.0 20.0 17.1 15.0 10.0 6.04 3.02 1.51 0.78 0.38 0.19 0.09 0.05

AFB1 AFB1

(A) (B)

able T 52

CHAPTER II. MATERIAL AND METHODS

II.2.3.3 IN VITRO TRANS-EPITHELIAL TRANSPORT

NCI-N87 and Caco-2 cells were seeded at 1×105 cells/cm2 in 24 mm 6 well Transwell with pore size of 0.4 μm and growth area of 4.67 cm2 (Corning incorporated, NY, USA). During differentiation the medium was changed every two days and the system used on its 28th post-confluence day to perform the transport assay. For transport experiments, the NCI-N87 cell medium was replaced by acidified HBSS (pH 3) and HBSS (pH 7.4) in the apical and basolateral compartments, respectively; while for Caco-2 cells, the apical compartment was replaced by HBSS with 25 mM HEPES and the basolateral compartment by HBSS with 25 mM HEPES and 0.5% (w/v) BSA. BSA was added to prevent binding of compounds to plastic (231) and simulate serum proteins. Mycotoxins at the initial concentrations of 1.69 µM (DON), 1.60 µM (AFB1), 0.69 µM (FB1), and 1.24 µM (OTA) (i.e. in equal mass ratio of approximately 500 µg L-1 each) were prepared in HBSS medium and introduced in the apical or basolateral compartments isolated or in mixture. After 15, 30, 60, 120 and 180 min, 200 μL aliquots were taken from the apical or basolateral compartments and replaced the same amount (200 μL) by fresh medium. The initial mycotoxins concentrations were also determined in the treatment solution as well as the final concentrations in both apical or basolateral compartments. These high concentrations are not expected to be found neither in food nor along the gastrointestinal tract considering the low concentrations of these toxins reported in literature and the strict regulations used to control maximum contents of these mycotoxins in food (309).

CHAPTER II. MATERIAL AND METHODS

II.3 CHEMICAL ANALYSES

II.3.1 NUTRITIONAL ANALYSIS

II.3.1.1 PROXIMATE COMPOSITION ANALYSIS

The proximate composition of each sample was determined according to the Association of Official Analytical Chemists (AOAC) (310): moisture by oven drying at 105 ºC; total fat and protein by Soxhlet and Kjeldahl procedures, respectively, and the ashes by heating in a muffle furnace at 500 ºC. The true retention (TR) percentage for each nutrient was calculated using the following formula: TR = (content per g of cooked food / content per g of raw food) x cooking yield (307). In the case of antibacterial and coccidiostats drugs, and mycotoxins, only the moisture and true retention percentage were analysed.

II.3.1.2 FREE AMINO ACID ANALYSIS

The extraction of free amino acids was made according to Dashdorj and colleagues (311) using 200 mg of meat sample homogenized with 1 mL of HCl 0.01N, left 30 min to free the amino acids, and centrifuged (17000 g, 1 min, 4 °C). Then, 300 µL of supernatant or 300 µL of digested sample were homogenized with 700 µL of acetonitrile to precipitate protein or large peptides, vortex, and centrifuged again (17000 g, 1 min, 4 °C). The supernatant (100 µL) was used for the analysis of free amino acids using the EZ:Faast Amino Acid Analysis Kit available for GC-FID (Phenomenex, USA) (312). Valine (Val), leucine (Leu), isoleucine (Ile), threonine (Thr), methione (Met), lysine (Lys), histidine (His), phenylalanine (Phe), tyrosine (Tyr), and tryptophan (Trp) were selected concerning their higher susceptibility to oxidation. Cystine was also measured as a product of cysteine oxidation. The free amino acid analyses were only performed in raw, oven cooked, and oven digested samples with/without the addition of oregano or beer.

II.3.1.3 TOTAL FATTY ACID ANALYSIS

The lipid extraction was performed according to Cruz and colleagues (313). Briefly, 50 µL of internal standard solution (10 mg/mL of undecanoin (C11:0 triglyceride) and 20 mg/mL of methyl tridecanoate (C13:0 methyl ester)) and 50 µL of BHT (10 mg/mL) were added to 800 mg of raw or cooked meat. In the case of digested samples, 50 µL of a diluted standard solution (1:100) were added to 2 mL of digested sample. Then 3 mL of isopropanol were 54

CHAPTER II. MATERIAL AND METHODS added to precipitate the protein followed by the addition of cyclohexane (4 mL); the raw and cooked meat samples were left overnight to allow the extraction of lipids. A shorter extraction time (30 min) was used in digested samples as lipids were more exposed. Then, 3 mL of NaCl aqueous solution (1%) was added to allow the separation of the upper organic phase containing the lipids, centrifuged (3000 g, 5 min), separated and evaporated to dryness under a gentle stream of nitrogen (Stuart®, Staffordshire, USA) and prepared for fatty acid analysis based on Dalziel and Colleagues (314). Briefly, the extracted fat was dissolved in 200 µL of heptane and 3 mL of 0.4 M H2SO4 in methanol was added. The mixture was incubated at 60 °C over 3h, with agitation at each 15 min. After cooling to room temperature, 5 mL of 0.43M K2CO3 was added to neutralize the solution. Then, 1.8 mL of heptane was added to fix the volume to 2 mL, vortex, and centrifuged (2,000 rpm, 2 min).

The heptane layer was transferred to a new tube and Na2SO4 was added to remove water residues prior to analysis. Then, 1 mL of extract was transferred to a 2 mL vial and fatty acids were determined by GC (Chrompack CP-9001 model, Netherlands) with flame ionization detection (FID). Fatty acid separation was carried out on a Select FAME (100 m x 0.25 mm x 0.25 µm) column (Agilent, USA) using helium as carrier gas (pressure of 190 kPa), and temperature of injector and detector set at 250 and 260 ºC, respectively. Data were processed by the CP Maitre chromatography data system program (Chrompack International B. V., Middelburg, Netherlands, version 2.5). The total fatty acid analyses were only performed in raw, oven cooked, and oven digested samples with/without the addition of oregano or beer.

II.3.2 COOKING-INDUCED HAZARDS

II.3.2.1 MALONDIALDEHYDE ANALYSIS (TBARS)

TBARS measurement as equivalents of free MDA in meat and digests was adapted from the methods described by Mendes and Colleagues (315) and Van Hecke and Colleagues (172). Firstly, 150 mg of samples, 400 µL of digested or 400 µL of standard solution were measured in a 1.5 mL microtube and the volume was fixed to 1 mL with a solution of TCA 7.5 % (w/v), vortexed, and sonicated during 5 min to allow release of MDA from matrix and to precipitate the protein. The samples were centrifuged (3000 g, 5 min) and the supernatant collected to a new tube; a second TCA precipitation was made, centrifuged, and the supernatant transferred to the tube with the previous supernatant. Then, 500 µL of supernatant was added to react with 500 µL of thiobarbituric acid (40 mM, prepared in acetic acid glacial), vortexed, and heated in a water bath at 90 ºC for 45 min. Then, the microtubes were cooled in an ice bath for 10 min and the absorbance read at 532 nm. MDA

CHAPTER II. MATERIAL AND METHODS

quantification was made using a standard curve with 1,1,3,3-tetramethoxypropane (TEP) (0.2 – 12.8 µmol) and the results were expressed as nmol MDA per g of sample.

II.3.2.2 4-HYDROXY-2-NONENAL AND HEXANAL

Determination of free HNE and HEX in meat and digests was adapted from the methods described by Mezzar and Colleagues (316) and Van Hecke and Colleagues (172). Briefly, 400 mg of samples, 400 µL of digested or 400 µL of standard solution were weighted in 4 mL tubes and 40 nmol of nonanal (NON, internal standard) was added. After addition of 400 µL of methanol (protein precipitation), an equal volume (400 µL) of derivatization medium (containing 3.6 M ammonium acetate, 100 mM CHD, 42% (v/v) acetic acid) was added, vortex, and heated in an oven at 70 ºC for 60 min. After the reaction time, the samples were cooled and 800 µL of methanol was added, vortex, and 1 mL of reaction solution was transferred to a 1.5 mL microtube and centrifuged (10000 g, 2 min, 4ºC). Then, 500 µL of supernatant was diluted with methanol (1:1) to ensure that all protein was precipitated, centrifuged (10000 g, 2 min, 4 ºC) and 10 µL of supernatant was injected on an HPLC system (Jasco LC PU-4180 Net II/ADC, Japan) with an auto-sampler (Jasco AS – 4050, Japan), and using a Kinetex C18 column (100 x 3.0 mm, 2.6 µm, 100 Å), with the column kept at 25 ºC (oven, ECOM Eco2000, Czech Republic). The derivatized compounds were eluted using a flow rate of 0.4 mL/min of water (eluent A) and methanol (eluent B) as mobile phases, with the following gradient: 50% B in the first 5 min, increasing to 100% B in 5 min, holding this conditions 2 min, and then decreasing to 0% in 1 min, and restoring the initial conditions at minute 13, followed by 8 min of stabilization (total run time of 21 min). HNE, HEX and NON were detected by fluorescence (FLD, Jasco FP – 4025, Japan) using an excitation/emission wavelengths of 380 nm and 446 nm, respectively. ChromNAV control center v2 – JASCO Chromatography Data Station was used to analyse the data. HNE and

HEX were quantified by plotting a standard curve (ACompound/ANonanal vs CCompound(nmol) with a linear range between 2 and 40 nmol of HNE or HEX) and expressed as nmol of HNE or HEX per g of sample.

II.3.2.3 CARBONYLS

The carbonyl content was determined by derivatization of protein carbonyls with 2,4- dinitrophenylhydrazine (DNPH) leading to the formation of stable dinitrophenylhydrazones adducts detectable at 370 nm (317). Briefly, one gram of cooked sample was homogenized overnight with 10 mL of 20 mM sodium phosphate buffer containing 0.6 M NaCl (pH 6.5) 56

CHAPTER II. MATERIAL AND METHODS using an orbital homogenizer to allow the dissolution of protein. In the case of raw samples, a faster dissolution was observed only requiring 30 min of homogenization. Then, two equal aliquots of 300 µL of each sample were taken to 2 mL microtubes and 1 mL of cold 10% (w/v) TCA was added to precipitate the proteins followed by a centrifugation step (4500 g, 3 min, 4 ºC). One pellet was treated with 1 mL 2N HCl (protein concentration measurement) and the other with an equal volume of 0.2% (w/v) DNPH in 2N HCl (carbonyl concentration measurement). Both samples were incubated for 1 h at room temperature, and vortex at each 15 min. After reaction, the samples were precipitated with 10% (w/v) TCA (1 mL), centrifuged (5400 g, 3min, 4 ºC) and the pellet washed twice with 1 mL ethanol:ethyl acetate (1:1, v/v). The pellets were carefully drained and then dissolved with 1 mL of 20 mM sodium phosphate buffer containing 6 M guanidine HCl (pH 6.5), centrifuged (5400 g, 2 min, 4ºC (to remove insoluble fragments)) and the absorbance of supernatant was measured at 370 nm. Carbonyl content was calculated using the molar absorption coefficient of 22000 M-1 cm-1 at 370 nm for protein hydrazones and the protein content was measured using the Bicinchoninic Acid (BCA) assay. Carbonyl content was expressed as nmol carbonyl per mg of protein.

II.3.2.4 SCHIFF BASES

The fluorescence emission of Schiff bases was assessed according to Hu and Colleagues (318). Briefly, meat (1 g) was homogenized for 30 s with 5 mL of phosphate buffer solution (20 mM, NaCl 0.6 M, pH 6.5). Then 2 mL of extract or 8 mL of digest was diluted with 8 mL of solvent (dichloromethane: ethanol (2:1 v/v)) and vortexed for 30 s. After centrifugation (4000 g for 10 min) the upper phase was collected and 200 µL of supernatant transferred to a 96-well plate to measure the fluorescence intensities (FI). The emission spectra were recorded from 390 to 600 nm with the excitation wavelength set at 360 nm (Cytation 3Cell Imaging Multi-Mode Reader, BioTek, USA). All measurements were made in triplicate and the fluorescence intensities expressed in arbitrary units.

II.3.3 ANTIBIOTICS AND COCCIDIOSTATS DRUGS

II.3.3.1 EXTRACTION FOR RAW AND COOKED MEAT

The ACDs extraction in chicken meat samples (raw and cooked) was performed according to Bittencourt and colleagues (319) and Moretti and colleagues (320) with few modifications. Thus 3.0 g of homogenized raw or cooked chicken meat was weighted in a 15 mL polypropylene centrifuge tube and a fixed concentration of IS (100 µg/kg) was added.

CHAPTER II. MATERIAL AND METHODS

For transformation products study a higher concentration of IS was used (500 µg/kg). After leaving the samples overnight for equilibration, 6 mL of ACN:water (80:20 v/v) acidified with 1% acetic acid plus 400 µL of EDTA (0.1M) were added. EDTA was used because TCs complex with bivalent cations (321). The mixture was vortexed, homogenized over 30 min using the overhead shaker, and then submitted to an ultrasound bath (10min). Then, the samples were centrifuged at 5000 rpm for 5 min at room temperature, and 1 mL of supernatant was filtered (0.45 µm nylon syringe filter) and transferred to the 2 mL glass vials for further analysis in the LC systems. Each sample was extracted twice.

II.3.3.2 EXTRACTION FOR BIOACCESSIBLE FRACTIONS

To 1 mL of digested samples, a fixed concentration of internal standards (500 µg/kg) was added. The bioaccessible fraction was analysed in two ways: 1) The bioaccessible fractions were diluted in MeOH (1:5 v/v) to allow protein precipitation, homogenized, and directly analysed (dilute and shoot method); or 2) a QuEChERS (Quick, Easy, Cheap, Effective, Rugged, and Safe) extraction was followed: 1 mL of digested samples was mixed with 130 µL EDTA and 1 mL of ACN acidified with 1% formic acid. Then 200 mg of NaCl and 300 mg MgSO4 were added and rapidly vortexed for 10 s to avoid salt agglomeration. The ACDs remaining in the non-bioaccessible fraction were extracted following the above mentioned extraction method used for solid samples with solvent volumes adapted to the remaining amount of sample. Then, the samples were centrifuged at 5000 rpm for 5 min at room temperature, and 1 mL of supernatant was filtered (0.45 µm nylon syringe filter) and transferred to the 2 mL glass vials for further analysis in the LC-MS systems. Each sample was extracted twice.

II.3.3.3 CHROMATOGRAPHIC ANALYSES

The analysis of ACDs was performed using two distinct equipment, the ultra-high liquid chromatography coupled to a triple quadrupole mass spectrometer (UHPLC-QqQ-MS/MS) to quantify the ACDs in raw, cooked and digested samples; and the ultra-high liquid chromatography coupled to orbitrap mass spectrometer (UHPLC-Orbitrap-MS) in an attempt to identify possible degradation/transformation products that could help understand the fate of some ACDs after cooking and in vitro digestion procedures.

CHAPTER II. MATERIAL AND METHODS

II.3.3.3.1 UHPLC-QQQ-MS/MS

The ACDs analysis was performed using an Agilent series 1290 RRLC instrument (Santa Clara, CA, USA) coupled to an Agilent triple quadrupole mass spectrometer (6460 A) with a Jet Stream electronic spray ionization (ESI) source (G1958-65138) equipped with the MassHunter (Agilent) software for data processing. The UHPLC conditions were adapted from Gomez-Perez and colleagues (322). The chromatographic separation was performed using a Zorbax Eclipse Plus C18 column (1.8 µm, 100 x 2.1 mm) by means of acidified water (0.1% formic acid) and MeOH as mobile phases A and B, respectively. The chromatographic run was made at a flow rate of 0.3 mL/min with a total time of 12 min. The elution gradient was as follows: 95% A in the first minute (0 – 1 min), decreasing its content to 0% in 6 min (1-7 min), holding this condition 3 min (0% A, 7-10 min), and returning to the initial conditions (95% A) in 0.5 min (10-10.5 min). The injection volume was 5 µL with the autosampler kept at 11 ºC. All ACDs, except DNC, were ionized in positive ESI mode. All of them were detected using dynamic multiple reaction monitoring (dynamic-MRM) mode. Source gas temperature and flow were 325 ºC and 5 L/min, respectively. Sheath gas temperature and flow were 400 ºC and 11 L/min, respectively. Capillary and nozzle voltages were 3500 V and 500 V, respectively. The optimized MS/MS parameters are available in Table II.2.

CHAPTER II. MATERIAL AND METHODS

- -

70/29 80/60 67/17 80/29 65/29 55/40 85/13 65/29 70/37 55/29 75/40 60/40 90/40

CV/CE 113/29 155/29 100/40 160/29 100/15

- -

3rd

251.1>92.1 311.1>92.1

749.1>154.0 787.5>279.2 366.1>114.0 291.2>123.1 362.2>261.1 332.1>231.0 360.2>245.1 445.2>154.0 461.2>201.0 254.1>156.0 465.1>154.0 445.2>154.0 916.5>101.1 837.5>116.1 454.0>295.1 334.1>111.0

67/9 65/9 55/9

70/17 80/50 80/25 85/25 65/21 90/17 70/33 55/25 75/40 60/17 90/25

CV/CE 113/25 155/21 100/29 160/17 100/25 120/25

2nd

MRM(m/z) transitions

749.1>462.0 787.5>531.4 366.1>208.0 251.1>108.0 291.2>261.1 362.2>344.1 332.1>294.1 360.2>316.1 445.2>427.1 461.2>443.1 254.1>108.0 465.1>430.0 281.1>156.0 311.1>108.0 445.2>410.0 916.5>145.1 837.5>679.4 454.0>196.1 334.1>138.0 301.1>107.0

collision energy (eV). (eV). energy collision

67/5

70/21 80/60 80/25 65/17 55/17 85/29 65/13 90/33 70/21 55/17 75/40 60/37 90/21

CV/CE 113/13 155/17 100/21 160/21 100/10 120/10

1st

MS/MS analysis drugs. coccidiostats and of antibacterial MS/MS

254.1>92.1 281.1>92.1

749.1>444.0 787.5>431.3 366.1>349.1 251.1>156.0 291.2>230.1 362.2>318.1 332.1>314.1 360.2>342.1 445.2>410.0 461.2>426.0 465.1>448.0 311.1>156.0 445.2>428.1 916.5>174.1 837.5>158.1 454.0>160.2 334.1>155.0 301.1>137.0

fragmentor voltage (V); (V); CE fragmentorvoltage

3.5 4.6 4.6 6.1

5.23 8.96 3.68 4.21 4.43 4.61 4.63 4.86 4.89 4.95 5.52 5.83 6.54 6.65 6.69 7.33

MS/MS parameters for UHPLC the MS/MS

CIP

TYL

SDZ

CTC

TMP ENR OTC

NAR SMX DNC

AMX

–

SDMX

FLU(IS)

ROX (IS) ROX

ROB (IS) ROB

OFLO (IS) OFLO

DEME(IS)

SMMX(IS)

transition used for quantification; CV quantification; used for transition

–

60 TableII.

CHAPTER II. MATERIAL AND METHODS

II.3.3.3.2 UHPLC-ORBITRAP-MS

In other to study the formation of degradation products derived from cooking of contaminated chicken breast burgers, the samples (non-spiked and spiked) were also analysed by high-resolution mass spectrometry using a Thermo Fisher Scientific TranscendTM 600 LC (San Jose, CA, USA) apparatus coupled to a single mass spectrometer Orbitrap Thermo Scientific (ExactiveTM, Thermo Fisher Scientific, Bremen, Germany) an electrospray interface (ESI) (HESI-II, Thermo Fisher Scientific, San Jose, CA, USA) in positive and negative mode. The chromatographic separation was performed using an identical column as used for the UHPLC-QqQ-MS/MS analysis, as well as the same mobile phases, but using a lower flow rate (0.2 mL/min) and the following gradient: initial conditions were 95% A, changing to 100% B in 1 minute and keeping this conditions for 7 min (1-8 min), then increasing A to 100% in 4 min (8-12) with a total time of run of 14 min. The injection volume was 10 µL. The MS acquisition was performed using two alternating functions: 1) full scan, ESI+ or ESI-, without fragmentation (with the higher collisional dissociation (HCD) collision cell switched off, the mass resolving power 25000 FWHM, and scan time 0.25 s); or 2) all ions fragmentation, ESI+ or ESI-, with fragmentation (HCD on, collision energy 30 eV, the mass resolving power 10000 FWHM, and scan time 0.10 s). A mass range in the full scan experiment was set at m/z 50-1000. ESI parameters were as follows: spray, skimmer, capillary, and tube lens voltage were of 4kV, 18V, 35V, and 95V, respectively; sheath and auxiliary gas (N2, >95%) were 35 and 10 (adimensional), respectively; and the heater and capillary temperature were of 30 ºC and 300 ºC, respectively. The Orbitrap analyser was accurately calibrated using a mixture of acetic acid, caffeine, Met-Arg-Phe-Ala-acetate salt and Ultramark 1621 (ProteoMass LTQ/FT-hybrid ESI positive), and a mixture of acetic acid, sodium dodecyl sulfate, taurocholic acid sodium salt hydrat and Ultramark 1621 (fluorinated phosphazines) (ProteoMass LTQ/FT-Hybrid ESI negative), from ThermoFisher (Waltham, MA, USA). Table II.3 shows the UHPLC-Orbitrap- MS parameters for ACDs analysis. The analysis of possible degradation products of ACDs was carried out using Compound Discover® or MassChemSite® Softwares. The raw files obtained from LC-Orbitrap-MS analysis of each sample were processed in order to look for unknown compounds related to ACDs. Compound Discover software allows to use a workflow of “degradants and unknown compounds” processing to search for each possible metabolic pathways from different transformations like, dehydration, reduction, methylation, S-dealkylation, sulfonation, desaturation and oxidation by known previously the structure of the parent

CHAPTER II. MATERIAL AND METHODS

compound (i.e of each ACD). In addition, ChemSpider and m/z Cloud databases were used to identify potential compounds. When the processing has been carried out, results were filtered according to intensity of the signal (higher than 1x104) and subtracting background (a blank sample of each cooking condition and raw controls were processed and used to eliminate possible false positives in the samples). When the data were filtered, a further study was developed, and all potential compounds were studied in order to identify potential transformation products. However, no reliable degradation products were found using this software. Therefore another software tool, MassChemSite®, was used. This software provided the reaction as well as the mass spectrum parent and the possible metabolite with the matches and mismatches for the full scan spectra and the HCD mass spectra. This software compared the different ions and searched for possible common ions, in order to corroborate if the ions came from the same family, due to the fact that in most cases the parent compound has common fragments with its transformation products.

Table II.3 - UHPLC-Orbitrap-MS parameters for ACDs analysis. Mass error Molecular formula RT Ion Predicted mass (ppm) + SDZ C10H10N4O2S 5.45 [M+H] 251.05969 -0.131 + AMX C16H19N3O5S 5.16 [M+H] 366.11209 0.743 + TMP C14H18N4O3 6.39 [M+H] 291.14505 -0.402 + OFLO C18H20FN3O4 6.65 [M+H] 362.15082 -0.665 + TC C22H24N2O8 6.84 [M+H] 445.16061 0.152 + CIP C17H18FN3O3 6.87 [M+H] 332.14032 -0.53 + ENR C19H22FN3O3 6.89 [M+H] 360.17151 0.038 + OTC C22H24N2O9 6.95 [M+H] 461.15561 0.332 + SMX C10H11N3O3S 7.27 [M+H] 254.05922 -0.663 + DEME C21H21ClN2O8 7.29 [M+H] 465.10620 0.603 + SMMX C11H12N4O3S 7.37 [M+H] 291.07025 -0.133 + CTC C22H23ClN2O8 7.75 [M+H] 479.12201 0.919 + SDMX C12H14N4O4S 8.12 [M+H] 311.08081 -0.135 + DC C22H24N2O8 8.31 [M+H] 445.16101 1.051 + TYL C46H77NO17 8.87 [M+H] 916.52740 1.058 + ROX C41H76N2O15 9.47 [M+H] 837.53302 1.397 + ROB C15H13Cl2N5 9.63 [M+H] 334.06195 -0.381 + FLU C19H16ClFN3O5S 9.66 [M+H] 454.06329 0.064 - DNC C19H18N6O6 10.65 [M-H] 301.05807 -0.152 + NAR C43H72O11 13.27 [M+Na] 787.49713 0.571

CHAPTER II. MATERIAL AND METHODS

II.3.4 MYCOTOXINS

II.3.4.1 EXTRACTION FOR RAW AND COOKED MEAT, AND HERBS

Mycotoxins extraction from chicken breast and herbs was performed using a QuEChERS-based approach according to Yogendrarajah and colleagues (323) with few modifications. Before extraction, the herbs were homogenized at 1000 rpm for 10 s (Grindomix GM 200, Retsch GmbH, Germany), and cooked chicken breast muscle was homogenized for three intervals of 20 s at 7500 rpm, with 5 s of pause between each homogenization interval (Precellys Evolution, Bertin Technologies). One gram of sample was weighted in a 50 mL polypropylene centrifuge tube and a fixed concentration of internal standard d5-OTA (20 µg/L) was added. After leaving the samples overnight for equilibration, 5 mL of water were added and shacked for 30 min. Thereafter, 5 mL of extraction solvent

(acidified ACN with 1% formic acid v/v) was added along with 2.0 g of MgSO4 anhydrous salt and 1.0 g of NaCl and the tube was immediately vortexed for 30 s to prevent agglomeration of the salts. The tubes were then centrifuged at 4000 g for 7 min to induce phase separation and mycotoxins partitioning. The organic phase was transferred to a 4 mL vial, and, in the case of chicken breast muscle, no clean-up procedure was used, while, for herbs, 1 mL of extract was purified through a C18 column (0.1 g) (previously activated with

1 mL H2O and 1 mL MeOH). Then, 1 mL of the chicken breast muscle extract and the cleaned-up herb extract were evaporated to dryness under a stream of nitrogen (Stuart®, Staffordshire, USA). The final extract was reconstituted in 250 µL of mobile phase B (methanol: water: acetic acid (97:2:1) with 5mM ammonium acetate) and transferred to a 2 mL glass vial for LC-MS/MS analysis. Each sample was injected twice.

II.3.4.2 EXTRACTION FOR DIGESTED AND TRANSPORTED SAMPLES

Two hundred microliters (200 µL) of digested sample (Section II.2.2.2) or aliquot took during transport assay (Section II.2.3.3) were transferred into a conic microtube and a fixed concentration of d5-OTA (20 µg/L) was added. Thereafter, 200 µL acidified ACN with 1% formic acid (v/v) was added along with 70 mg of MgSO4 anhydrous salt and 10 mg of NaCl and the tube was immediately vortexed for 10 s to prevent agglomeration of the salts. The tubes were then centrifuged at 17,000 g for 5 min to induce phase separation and mycotoxins partitioning. The organic phase was transferred to a 2 mL vial, evaporated to dryness under a stream of nitrogen (Stuart®, Staffordshire, USA), and finally reconstituted in 60 µL of mobile phase B and analysed by LC-MS/MS. Each sample was injected twice.

CHAPTER II. MATERIAL AND METHODS

II.3.4.3 CHROMATOGRAPHIC ANALYSIS

The analysis of the target mycotoxins was performed on a Waters 2695 HPLC system (Water, Milford, MA, USA) coupled to a Micromass Quattro micro API™ triple quadrupole detector (Waters, Manchester, UK), equipped with the MassLynx 4.1 software for data processing. The HPLC conditions were according to Cunha and colleagues (324). The chromatographic separation was achieved using a Kinetex® Phenomenex® C18 column (2.6 μm, 150 mm x 4.60 mm (i.d.)) with a pre-column from Phenomenex (C18, 4 x 3.0 mm (i.d.), Torrance, CA, USA) and the column kept at 35 ºC. The mobile phase consisted of a ternary mixtures of mobile phase A (water/methanol/acetic acid, 94:5:1 v/v and 5 mM ammonium acetate) and mobile phase B (methanol/water/acetic acid, 97:2:1 v/v and 5 mM ammonium acetate), at a flow rate of 0.3 mL/min, and volume injection was 20 µL. The solvent gradient program was as follows: (1) 0-7.0 min, 95% A and 5% B; (2) 7.0-11.0 min, 35% A and 65% B; (3) 11.0-13.0, 25% A and 75% B min; (4) 13.0-15.0, 0% A and 100% B min; (5) 15.0-24.0 95% A and 5 % B min; and (6) 24.0-27.0 min, 95% A and 5% B min. MS/MS acquisition was operated in positive-ion mode with multiple reaction monitoring (MRM). The optimized MS parameters were as follows: capillary voltage, 3.00 kV; source temperature, 150 ºC; desolvation temperature, 350 ºC; desolvation gas and cone gas flow, 350 and 50 L/h, respectively. High purity nitrogen (≥99.999%, Gasin, Portugal) and argon (≥99.995%, Gasin, Portugal) were used as cone and collision gas, respectively. Dwell times of 0.2 s/scan were selected. For each analyte, two transitions were selected for identification and the corresponding cone voltage and collision energy were optimized for maximum intensity. The optimized MS/MS parameters for target mycotoxins are listed in Table II.4.

CHAPTER II. MATERIAL AND METHODS

Table II.4 - MS/MS parameters for the studied mycotoxins.

Retention time MRM transition (m/z) CV (V) CE (eV) Mycotoxins (min) QIT (m/z) CIT (m/z) QIT CIT QIT CIT

DON 9.73 297>203.3 297>249 22 20 13 11 AFG2 11.86 330.8>313.1 330.8>245.3 35 35 24 30 AFG1 12.28 329>243 329>311.2 35 35 30 30 AFB2 12.82 315>259.2 315>287.3 45 45 33 35 AFB1 13.26 313>241.2 313>285.2 45 45 30 30 FB1 15.12 722.5>334.2 722.5>352.4 46 44 40 36 T2 17.08 484>214.9 484>245.2 21 23 18 15 FB2 18.15 706.3>336.11 706.3>318.4 42 40 38 36 OTA-d5 (IS) 18.54 409>239.4 409>358.1 32 32 22 22 OTA 18.6 404>239.1 404>358.1 30 28 20 16 ZEN 18.8 319.2>187 319.2>283.3 20 20 18 16 Abbreviations: AFB1 - aflatoxin B1; AFB2 – aflatoxin B2; AFG1 - aflatoxin G1; AFG2 - aflatoxin G2; CE - Collision energy; CIT - Confirmation ion transition; CV - Cone voltage; DON - deoxynivalenol; FB1 - fumonisin B1; FB2 - fumonisin B2; OTA - ochratoxin A; OTA-d5 - Ochratoxin A-d5; QIT - Quantification ion transition; T2 - T-2 toxin; ZEN - zearalenone.

II.3.5 QUALITY CONTROL/QUALITY ASSURANCE

This section includes two different quality control measurements: 1) the quality control used during development and validation of chromatographic methods (Section II.3.5.1); and ii) quality control measurement of monolayers integrity during the transport assays to ensure that the experiment was not compromised (Section II.3.5.2).

II.3.5.1 QUALITY CONTROL OF CHROMATOGRAPHIC METHODS

Prior to analysis, vials were baked for 1h at 300 ºC and washed with acetone to remove possible contamination. Throughout all chromatographic analysis herein performed at least two independent blanks were extracted within each batch of samples and injected randomly. All experiments had three independent replicates of each sample and each replicate was extracted twice giving a total of six values per sample. Also, analytical blanks were performed at each 6 injections to clean/control the system. The chromatographic methods were evaluated by considering an in-house validation approach evaluating the linearity, accuracy, and precision to perform reliable quantitative results according to EU guidelines and the International Conference on Harmonization

CHAPTER II. MATERIAL AND METHODS

(ICH) recommendations (325, 326). All chromatographic analyses used internal standards to aid quantification of target analytes and correct possible losses of sample during preparation. Thus, the quantification of target compounds was made by plotting calibration curves as follows: Area target compound / Area respective internal standard versus Concentration target compound. HNE and HEX calibration curves were made using an isotonic solution as solvent (0.9% (w/v) sodium chloride solution) through the linear range of each compound (see Section II.3.2.2). In the case of ACDs and mycotoxins analysis, the matrix effect was evaluated in raw meat for all the compounds included in this study and it was expressed as

푆푙표푝푒 푖푛 푚푎푡푟푖푥 푀푎푡푟푖푥 푒푓푓푒푐푡 (%) = − 1 푥 100 푆푙표푝푒 푖푛 푠표푙푣푒푛푡

For those compounds showing significant matrix effect (>20% or <-20%) suggesting enhancement or suppression of analytical response, the quantification was performed using matrix-matched calibration curves for each analyte, using at least five calibration points, through the linear range of the compounds, as suggested by SANTE guidelines (327). ACDs quantification was performed in a range of 10 to 1500 µg/kg in raw and cooked meat and 75 to 1500 µg/L in bioaccessible digested fractions (QuEChERS method), while mycotoxins quantification was performed in the range 3.0 to 200 µg/kg in raw/cooked meat and herbs and 1.0 to 100 µg/L in the bioaccessible digested fractions. Accuracy was measured through recovery studies and precision was determined by repeatability (intra-day precision) and intermediate precision (inter-day precision) of samples spiked, at two concentration levels (ACDs) or three concentration levels (mycotoxins), using at least three replicates, in the first day, and two on the second day of analysis. Different approaches were used to determine LODs and LOQs. LODs were only determined in mycotoxins analyses by successive analyses of chromatographic extracts of sample solutions spiked with decreasing amounts of the analytes until a signal-to-noise ratio 3:1 was reached. The LOQs were estimated differently depending on the analysis: i) In the case of HNE and HEX the limits of quantification (LOQs) were set as the lowest concentration point of calibration curve; ii) in the case of mycotoxins, these were determine the same way as LODs, but considering a signal-to-noise of 10:1; and iii) in case of ACDs, the estimation of LOQ was set as the minimum concentration that provides suitable trueness (recovery) and precision values according to SANTE guidelines (327) not taking into consideration only the S/N criterion, because this last one may lead to misunderstandings and even to underestimate LOQ values and provides unrealistic low 66

CHAPTER II. MATERIAL AND METHODS

LOQ values. For ACDs and mycotoxins experiments in meat and herbs, every sample for each matrix was analysed to check the presence of target compounds in the matrices in study, and thus used to correct the value obtained. Nevertheless, none of the tested matrices presented detectable amounts of target analytes.

II.3.5.2 QUALITY CONTROL OF MONOLAYERS FOR TRANSPORT ASSAYS

To perform transport assays at least two controls shall be taken into account:

1) Ensure that the concentration of the target compounds used for transport assay does not lead to cell death or membrane detachment. For that cytotoxic assays are performed as previously described in Section II.2.3.2.1; 2) Measure the trans-epithelial electrical resistance (TEER) of the monolayers before and after the experiment to ensure that the membrane was not compromised during the experiment. TEER values were determined at 37 ºC using the Millicell ERS-2 Voltohmmeter (Merck Millipore, Darmstadt, Germany) at the beginning and at the end of the experiment after monolayer washing with HBSS, to check the barrier integrity. The resistance, expressed as Ω cm2, was calculated by multiplying the cell monolayer resistance (Ω) by area of the filter (cm2). All transport assays were performed in triplicate. TEER values are strong indicators of integrity of cellular barriers and the measurement can be performed in real-time without damaging the cells and is generally based on measuring ohmic resistance between the apical and basolateral compartments. Also, fully differentiated cultures should exhibit TEER values between 500 and 1100 Ω cm-2 to be considered suitable for the experiment (328). Moreover, the mass balance recovery could also be used as a quality control of the experiment. The mass balance should be as close as possible to 100%, and values above 80% should give acceptable approximation of the apparent permeability value estimation (231). The mass balance was calculated using the equation:

Mass balance (%) =[(VR×CR(final))+(VD×CD(final))]/( VD(0)×CD(0))×100

where CR and CD are the concentrations on the receiver (R) and donor (D) sides of the monolayer at the beginning (0) and end (final) of the experiment, and V is used for each of the respective volumes.

CHAPTER II. MATERIAL AND METHODS

II.4 EXPOSURE ESTIMATION

II.4.1 BIOACCESSIBILITY ESTIMATION

The bioaccessibility estimation was only made for ACDs and mycotoxins, and not for cooking-induced hazards, because the later ones are formed during digestion meaning that their content increases after ingestion, whereas ACDs and mycotoxins are not formed during digestion, thus it is expected that their contents would never exceed the concentration found in samples before digestion. Thus, the gastric and/or intestinal percentages of bioaccessibility of ACDs and mycotoxins from digested cooked meat were calculated as follows:

푎푚표푢푛푡 푖푛 퐵푖표푎푐푐푒푠푠푖푏푙푒 푓푟푎푐푡푖표푛 (푛푔) Bioaccessibility (%) = 푥 100 푎푚표푢푛푡 푖푛 푐표표푘푒푑 푠푎푚푝푙푒푠 (푏푒푓표푟푒 푑푖푔푒푠푡푖표푛)(푛푔)

II.4.2 APPARENT PERMEABILITY AND FRACTION ABSORBED ESTIMATION

In the transport study of mycotoxins, the gastric and intestinal permeabilities of mycotoxins were determined using transport rates across NCI-N87 and Caco-2 cell monolayers, as described elsewhere (231). Permeability coefficient (Papp) at both directions (ApicalBasolateral (AB) and BasolateralApical (BA)) was calculated according to the following equation, described for experiments under non-sink conditions (201, 231):

1 1 푀 푀 −푃app퐴( + )푡 퐶R(푡) = + (퐶R0 − ) 푒 푉D 푉R 푉D+푉R 푉D+푉R

where CR (t) is the time-dependent mycotoxin concentration in the receiver compartment, M is the amount of mycotoxin in the system, VD and VR are the volumes of the donor and receiver compartment, respectively, CR0 is the concentration of the mycotoxin in the receiver compartment at the start of the time interval, A is the area of the filter, and t is the time from the start of the interval.

Papp was obtained from nonlinear regression, minimizing the sum of squared residuals

2 ∑(CR,i,obs - CR,i,calc) where CR,i,obs is the observed receiver concentration at the end of the interval and CR,i,calc is the corresponding concentration calculated according to the previous equation (329).

CHAPTER II. MATERIAL AND METHODS

Uptake (UR) and efflux (ER) ratios were calculated as the quotient of absorptive (to plasma) (Papp AB/Papp BA) and secretory (to intestinal lumen) (Papp BA/Papp AB) permeabilities, respectively.

Apical to basolateral Papp data on Caco-2 cells were used to estimate the calculated human fraction absorbed, FA (%), according to the nonlinear regression model described by Skolnik and colleagues (330) and Tavelin and colleagues (331):

100 FA (%) = 1+푒((−5.74−P푎푝푝푋)/0.39)) where 100 equals the minimum + maximum – minimum of % FA values constrained to 1 and 100 %; -5.74 is the log PappAB value at 50% of absorption in humans, PappX is the log

Papp AB for Caco-2 cells of mycotoxins obtained in the present study, and 0.39 is the slope that derived from the model fit.

II.4.3 TOXICOLOGICAL EVALUATION

The evaluation of mycotoxins toxicity isolated and combined was performed using the Compusyn software version 1.0 (ComboSyn Inc., Paramus, NJ, USA). This software was used to calculate the inhibitory concentration at 50% for combination index studies. This software calculates the IC50 using the median-effect equation of the mass-action law described by Chou and Talatay, considering the shape of the dose-response curve, either when m=1 hyperbolic, m>1 sigmoidal, m<1 flat sigmoidal (332):

log (푓푎/푓푢) = 푚푙표푔퐷 − 푚푙표푔퐷푚

where D is the dose, Dm is the dose required for 50% inhibition, fa is the fraction affected by the dose D, m is the coefficient of the sigmoidicity of the dose effect, and fu = 1 - fa. The effect of the mycotoxins binary combination on cells was analysed through the combination index (CI) theorem of Chou and Talatay (332) developed for quantification of synergism or antagonism for two drugs:

(퐷) (퐷) 퐶퐼 = 1 + 2 (퐷푚)1 (퐷푚)2

CHAPTER II. MATERIAL AND METHODS

where (Dm)1 and (Dm)2 are the doses of individual mycotoxins corresponding to the IC50; and, (D)1 and (D)2 are the doses of the two mycotoxins that combined inhibited cell growth to half (IC50). If the sum of these equations is around 1 (0.9 < CI < 1.1), the mycotoxins combination has an additive effect on cells. CI < 0.9 or CI > 1.1 mean synergism or antagonism effect, respectively. CI values were expressed as CI-Plot and isobologram graphs. For those combinations showing synergism, dose reduction indices (DRI) were calculated (277). DRI indicates how many folds the dose of each compound in the combination may be reduced to achieve the same toxic level compared with the doses of each compound alone. DRI values can be obtained from the reciprocal of CI equation:

(퐷푥)1 (퐷푥)2 (퐷푅퐼)1 = and (퐷푅퐼)2 = (퐷)1 (퐷)2

where (Dx)1 and (Dx)2 are the doses that the mycotoxins alone inhibited x % of cells, and D1 and D2 are the values that mycotoxins combined inhibited x % of cells.

CHAPTER II. MATERIAL AND METHODS

II.5 STATISTICAL ANALYSES

For all experiments, all dependent variables were tested for normal distribution using Shapiro–Wilk’s test. For normally distributed variables One-way ANOVA and ANCOVA analysis were applied, using Tukey’s or Tamhane’s T2 post hoc tests, depending on their homogeneity of variances. The variables not following normal distribution, Kruskal–Wallis test and Dunn’s post hoc test was applied for median comparison. In the case of mycotoxins, partial least squares (PLS) regression was used to study the relationships between percentage of mycotoxin after cooking or bioaccessibility (Y-matrix) and type of cooking applied (X-matrix) in terms of prediction of Y-variables from X-variables. The distribution of the eight types of cooking used on samples with 10 studied mycotoxins regarding the percentage of mycotoxin remained after cooking/bioaccessible was evaluated by Agglomerative Hierarchical Clustering (AHC) analysis and then displayed as a cluster heatmap. Therefore, Euclidean distance and Ward linkage were selected to establish the clusters (333). A 5% signiﬁcance level was considered for all statistical analyses. ANOVA and Kruskal-Wallis test, and PLS regression were performed using XLSTAT for Windows version 2016.02 (Addinsoft, Paris, France). Heatmap plots and AHC were carried out using the heatmap.2 function from the gplot package in R (http://cran.r- project.org/web/packages/gplots/index.html). All data are expressed as mean ± standard deviation (SD) or median (minimum – maximum) of three independent experiments and extracted twice (n=6). GraphPad Prism version 7.00 for Windows was used to build all graphs (Graphpad Software, La Jolla, California, USA).

CHAPTER III – RESULTS AND DISCUSSION

III.1 PROTEIN AND LIPIDS STABILITY

III.1.1 IMPACT OF CULINARY PRACTICES

III.1.2 LIPID AND PROTEIN OXIDATION THROUGH IN VITRO DIGESTION

III.1.3 FINAL REMARKS

III.2 ANTIBACTERIAL AND COCCIDIOSTATS DRUGS (ACDS)

III.2.1 METHOD DEVELOPMENT AND VALIDATION

III.2.2 IMPACT OF CULINARY PRACTICES

III.2.3 ACDS BEHAVIOUR TOWARD IN VITRO DIGESTION

III.2.4 EXPLORATORY AND STATISTICAL ANALYSIS

III.2.5 FINAL REMARKS

III.3 MYCOTOXINS

III.3.1 METHOD DEVELOPMENT AND VALIDATION

III.3.2 IMPACT OF CULINARY PRACTICES

III.3.3 MYCOTOXINS’ BEHAVIOUR TOWARD IN VITRO DIGESTION

III.3.4 GASTRIC AND INTESTINAL ABSORPTION OF MYCOTOXINS

III.3.5 TOXICOLOGICAL INTERACTION BETWEEN MYCOTOXINS

III.3.6 FINAL REMARKS

Parts of this chapter are available in the following publications:

Sobral MMC, Faria MA, Cunha SC, Miladinovic B, Ferreira IMPLVO. Transport of mycotoxins across human gastric NCI-N87 and intestinal Caco-2 cell models. Food and Chemical Toxicology- 2019. 131:110595.

Sobral MMC, Romero-Gonzalez R, Faria MA, Cunha SC, Ferreira IMPLVO, Garrido-Frenich A. Stability of antibacterial and cocciodiostat drugs on chicken meat burgers upon cooking and in vitro digestion. Food chemistry. 2020. 316:126367.

CHAPTER III. RESULTS AND DISCUSSION

III. 1 PROTEIN AND LIPIDS STABILITY

Meat thermal treatment, mandatory before consumption, triggers the generation of ROS and consequent oxidation of food constituents. In particular, PUFAs oxidation results in the formation of LOPs, such as MDA, HEX, and HNE. Amino acids are also oxidized during cooking, increasing protein denaturation and aggregation due to the formation of disulfide and dityrosine bridges, and leading to the formation of Schiff bases. The addition of spices and herbs has been shown to be effective against the formation of LOPs during food thermal treatment, as well as red wine polyphenols have also shown protective effects against oxidation (170). However, the influence of beer on protein and lipid oxidation remains unknown either after cooking or after in vitro gastrointestinal digestion. Moreover, the formation of these oxidation compounds seems to increase during in vitro digestion, mostly in the acidic environment of the gastric phase which has been described as a bioreactor of lipid peroxidation (172). The majority of studies regarding oxidation processes focus on beef meat oxidation, with the poultry meat receiving less attention, although being highly sensitive to oxidation owing to the unsaturation degree of their lipids. This chapter will provide knowledge on the formation of oxidation markers such as MDA, HNE, HEX, carbonyls and SB structures after two different home-cooking methods – oven or microwave cooking – with/without the addition of extra ingredients (oregano or beer), as well as after in vitro digestion of chicken burgers.

III.1.1 IMPACT OF CULINARY PRACTICES

III.1.1.1 LIPID OXIDATION

The formation of three lipid peroxidation-derived aldehydes – MDA, HNE, and HEX – and the fatty acid profile were used to evaluate the degree of lipid oxidation (LipOx) after oven cooking or microwaving, with or without oregano/beer. Raw samples showed an expected initial level of oxidation with MDA, HNE, and HEX values ranging from 2.34 to 7.91, 2.98 to 4.32, and 2.93 to 10.32 nmol/g of meat, respectively (Figure III.1) considering the oxidative reactions that occur in meat after slaughter. MDA values agree with previous reports concerning their contents in raw meat or fish (92, 102, 170, 318). The previous addition of herbs/beer did not alter the initial content of these aldehydes in raw samples suggesting that lipid oxidation was not observed during marinating overnight (~14h). On the contrary, cooking triggered the formation of the three aldehydes (p<0.05), contributing with a 18-fold and 22-fold increase of MDA, a 157-fold and a 84-fold increase

CHAPTER III. RESULTS AND DISCUSSION

of HEX after oven cooking and microwaving, respectively, and a 66-fold increase for HNE regardless the cooking method.

Figure III.1 – Effect of adding oregano or beer on the TBARS (expressed as MDA) (A), hexanal (HEX) (B), and 4-hydroxy-2-nonenal (HNE) (C) values before and after cooking (oven or microwave) chicken burgers. Values are means ± SD of three independent cooking experiments and each sample extracted twice (n=6). * p<0.05, ** p<0.001

The increased LOPs results from the damage of meat cellular structure promoted during cooking, which allows the free radicals to initiate the radical attack, promoting the peroxidation of unsaturated fatty acids and consequent formation of reactive aldehydes. Furthermore, cooking also releases the heme-iron from the porphyrin ring accelerating the oxidative deterioration (334). The type of heating - dry heat (oven) or radiation (microwaving) - seems not to influence MDA and HNE formation, while for HEX, lower values were found in microwaved samples (p<0.05). Contrary to our findings, Hu and colleagues (102) studied the influence of several cooking methods on fish fillets, reporting significant differences between roasting and microwaving, with higher TBARS values reported in microwaved samples. An increase of MDA, HNE, and HEX content was also recently reported by Van Hecke and colleagues (335) after thermally treating (water bath, 70 ºC, 70 min) chicken breast or thigh muscles.

CHAPTER III. RESULTS AND DISCUSSION

The type of ingredients added – oregano or beer – affected differently the formation of aldehydes during cooking. Adding beer to meat prior to oven cooking did not prevent MDA, HNE and HEX formation exhibiting similar amounts to oven cooked samples without ingredients. On the other hand, microwaving with beer reduced HNE and increased HEX formation. HNE and HEX are aldehydes deriving from the oxidation of ω-6 fatty acids, as the linoleic (18:2, ω-6) or arachidonic (20:4, ω-6), among others. Their formation differs on carbon where radical attack takes place: HNE results from radical attack and hydroperoxide formation at carbon 10, while HEX formation results from the radical attack at carbon 7. The increased values of HEX over HNE in the presence of beer may suggest that beer constituents favour radical attack at the double bond in carbon 7. Indeed, linoleic acid (C18:2) suffered a higher loss in cooked samples with beer (Figure III.2).

Figure III.2 – The variation of ω-3 and ω-6 fatty acids content in oven cooked samples using the raw samples as starting point content (raw content of each condition was subtracted to cooked samples) to evaluate their stability to cooking. Values are means ± SD of three independent cooking experiment and each sample extracted twice (n=6). The statistical analysis was performed within oven cooked or digested samples with different letters meaning significant differences within each group (p<0.05)

Furthermore, higher losses (p<0.05) in samples cooked with beer were observed for other ω-6 PUFAs, such as, linolenic acid (18:3), di-homo-y-linoleic acid (20:3), araquidonic acid (20:4), docosatetraenoic acid (22:4), and docosapentaenoic acid (22:5), as well as some ω-3 PUFAs like linolenic acid (18:3), stearidonic acid (18:4), eicosapentaenoic acid 77

CHAPTER III. RESULTS AND DISCUSSION

(20:5), and docosapentaenoic acid (22:5). Beer may have some characteristics that can promote oxidative conditions, such as the acidic pH (~4,5), the presence of alcohol (5% v/v), and the presence of ROS resulting from the brewing process (336, 337). In this sense, the radicals present in beer would further react with PUFAs from meat reducing their contents. Nonetheless, Pilsner beers are suggested as a good source of polyphenols and used as a mitigation strategy to prevent the formation of harmful compounds in cooked beef in both alcoholic and non-alcoholic beer marinades (115). Even so, according to Aron and colleagues (336) some polyphenols, such as flavan-3-ols, proanthocyanidins and flavonols may potentially promote ROS formation behaving as pro-oxidants able of modifying biological molecules such as lipids and proteins. The addition of oregano strongly prevented the formation of the three compounds, with no formation of MDA, HNE and HEX, in comparison with their raw controls (Figure III.1). These results agree with Van Hecke and colleagues (172) who reported a 50% reduction of MDA, no detection of HNE, and 87 to 94% reduction of HEX in beef cooked with oregano. These authors observed similar effects with the addition of other herbs such as basil, rosemary, and thyme. However, their study evaluated higher herb/meat proportions (0.5 and 1% w/w) than the one herein tested (0.2% w/w). This shows that lower oregano contents and therefore more realistic proportions can be used to effectively prevent MDA, HNE and HEX formation and preserve PUFAs. Indeed, higher contents of total PUFAs and ω-6 PUFAs (p<0.05) were observed in oven cooked samples with added oregano in comparison with oven cooked samples without ingredients (Table III.1). Oregano herbs are rich in phenolic compounds, namely flavonoids and phenolic acids, which may justify the preventive behaviour of these herbs towards LipOx (338). These compounds are involved in several antioxidant mechanisms including free radical scavenging, reducing agents capacity, metal chelation, and single oxygen quenching, among others (334). Definitely, natural ingredients have been introduced in industry to replace synthetic antioxidants and prevent meat oxidation during storage as well as during thermal treatments (339). Overall, cooking with oregano exhibited the lowest losses of the abovementioned PUFAs and the lowest formation of LOPs. Therefore, it should be added to meat prior to cooking to efficiently preserve ω-6 and ω-3 PUFAs, while beer addition should be avoided since it significantly affects the nutritional value of meat in terms of PUFAs content.

CHAPTER III. RESULTS AND DISCUSSION

III.1.1.2 PROTEIN OXIDATION

Severe cooking methods may also contribute to protein oxidation (ProtOx) through the oxidation of amino acids, including the essential ones, and change of protein conformation altering their final digestibility and consequently decrease the nutritional quality of cooked meat (74). Herein, the formation of carbonyls and SB structures was monitored as result of ProtOx, and the free amino acid content was evaluated as a direct consequence of their integrity loss through oxidation, although bearing in mind that free amino acids are in lower amounts than the amino acids from proteins (91). Val, Leu, Ile, Thr, Lys, His, Phe, Trp, Tyr, and Cystine were selected as the free amino acids more prone to oxidation according to literature (74, 91). Carbonylation is a non-specific irreversible reaction that takes place from an early stage of oxidative processes which makes carbonyls a widely acknowledged indicator of overall ProtOx (340). Microwaving and oven cooking did not promote carbonyl formation, which was not expected since literature shows an increase of carbonyls after thermal processing (87, 102, 179) (Figure III.3). However, recently studies from Van Hecke and colleagues (335) reported only a low increase of carbonyls after cooking chicken meat breast and thigh muscle.

Figure III.3 – Effect of adding oregano or beer on the carbonyl content before and after cooking (oven or microwave) chicken burgers. Values are means ± SD of three independent cooking experiments and each sample extracted twice (n=6). * p<0.05.

Similarly, oregano and beer did not influence carbonyl formation, exhibiting similar contents before and after cooking. Only oven cooking with beer exhibited higher carbonyls values (p<0.05), which agrees with lower contents of Thr, Tyr, Val, and Lys (p<0.05) observed in samples cooked with beer in comparison with those cooked with oregano (Figure III.4). 80

CHAPTER III. RESULTS AND DISCUSSION

Figure III.4 – The variation of free amino acids content in oven cooked samples using the raw samples as starting point content (raw content of each condition was subtracted to cooked samples) in order to investigate the stability of amino acids considered as susceptible to oxidation: (A) Polar, non-charged amino acids; (B) Aromatic amino acids, (C) Apolar amino acids, and (D) Polar, positively charged (91). Values are means ± SD of three independent cooking experiment, each sample injected three times (n=9). The statistical analysis was performed within oven cooked samples with different letters meaning significant differences within each group (p<0.05).

Moreover, the appearance of fluorescent structures between 390-470 nm in cooked meat that did not exist before cooking (raw sample with no ingredients) (Figure III.5) could suggest the interaction of carbonyls with other components of meat (e.g. oxidized aldehydes) and justify the lower carbonyl contents observed (341). Nonetheless, it should be mentioned that raw samples with oregano or beer also exhibited fluorescent structures between 390-470 nm probably from the increment of fluorescent structures from oregano/beer themselves or interactions/reactions that occurred between meat and added ingredients during marinating overnight (Figure III.5A). Thus, to properly analyse the impact of cooking on samples concerning the formation of SB, the variation of flourecence between cooked samples and respective raw controls was assessed (Figures III.5B and Figure III.5C). Undoubtedly, an increase of fluorescence intensity and the appearance of a second peak near 415 nm was observed after oven cooking and microwaving. This fluorescent peak has been linked with the formation of Maillard fluorescent products (342). The appearance 81

CHAPTER III. RESULTS AND DISCUSSION

of this second peak has already been reported after cooking fish and bovine meat (94, 102, 318).

Figure III.5 – Influence of adding oregano or beer on the emission fluorescence spectra (excitation at 360 nm) of raw (A) and cooked samples (B and C). Figures III.5B and III.5C show the variation of fluorescence in cooked samples using the fluorescence of raw samples as starting point content (fluorescence of raw samples of each condition was subtracted to cooked samples) in order to investigate the formation of SB during cooking. Values are means of three independent cooking experiments and each sample extracted twice (n=6).

The addition of beer similarly triggered SB formation as oven/microwave cooking themselves, while adding oregano before cooking seemed to prevent further formation of fluorescence structures, showing a lower increase of SB after cooking in comparison with the other cooking methods (Figure III.5B and Figure III.5C). The similarity of fluorescence behaviour concerning the type of cooking shows that the addition of ingredients is the most important factor on SB structures formation. The complexity of meat and the different reaction combinations between proteins and aldehydes may justify the countless fluorescent structures and different peaks observed (94). The direct measurement of fluorescence intensity on samples has been demonstrated to yield good indices of oxidation in biological materials such as ﬁsh and both raw and heated meat (342). Additionally, while

CHAPTER III. RESULTS AND DISCUSSION the values of some free amino acids decreased after cooking - free Met, Val, Leu, Ile, and Phe – as a result of their conversion into carbonyls and/or reaction with aldehydes via SB formation (102), it shall be mentioned that others - Thr, His, Tyr, and Trp – increased because thermal treatment of meat may also favour the proteolysis of amino acids (Figure III.4).

CHAPTER III. RESULTS AND DISCUSSION

III.1.2 LIPID AND PROTEIN OXIDATION THROUGH IN VITRO DIGESTION

The quality and health effects of nutrients that reach the bloodstream is dependent on the physiological process of meat digestion. After cooking meat, the remaining lipids, protein, and oxidized products are further exposed to pro-oxidant conditions along GIT, such as the low pH of gastric juice, dissolved oxygen content, and reactive species (e.g. metallic ions) existing in food bolus as well as their release from food components (176). Both static and dynamic in vitro digestion models have been used to study the influence of antioxidant compounds on LipOx and ProtOx during digestion of meat or fish (170, 172). Herein, the newest INFOGEST 2.0 static in vitro digestion model (191) was applied to ease the comparison with literature, as the majority of studies were conducted using static models, and also because the high number of samples required is only possible to process with a static methodology. Despite some differences observed between oven and microwave cooking, mostly in carbonyls determination, the type of ingredients – oregano or beer – revealed to have a prominent impact on LipOx and ProtOx, thus the in vitro digestion studies were performed using only the oven cooked samples, as we believe that the evolution after in vitro digestion of most of oxidation markers would be similar in both cooking methods and in vitro digestion assays are laborious and expensive. Figure III.6 shows the LOPs and carbonyls values measured after in vitro digestion of oven cooked samples, with a red line in each bar indicating the starting point before digestion (content found in cooked samples).

Figure III.6 - Effect of adding oregano or beer on the TBARS (expressed as MDA), hexanal (HEX), 4-hydroxy- 2-nonenal (HNE), and carbonyls values after in vitro digestion of oven cooked chicken burgers. The red lines represent the average content of each compound before digestion (content found in cooked samples). MDA, HNE, and HEX values are expressed as nmol of compound per g of digested meat. Values are means ± SD of three independent cooking/digestion experiments and each sample extracted twice (n=6). ** p<0.001

CHAPTER III. RESULTS AND DISCUSSION

MDA values increased after in vitro digestion, except in samples with oregano, while HEX and HNE values decreased after digestion of cooked meat samples without ingredients, remained equal in digested meat with beer, and increased in samples with oregano. Moreover, LOPs formation concerning the culinary practices evaluated could be ranked as oven without ingredients > oven with beer > oven with oregano regarding MDA values; and as oven without ingredients ≈ oven with beer > oven with oregano for HEX and HNE. No studies were found regarding the influence of Pilsner beer on LOPs formation during digestion, however incorporating melanoidins from dark beer showed lower MDA values after gastric digestion of grilled poultry meat in comparison with digested control Concerning fatty acid analysis, the digested samples without ingredients exhibited the lowest PUFAs content (1.14±0.05 g/100 g d.w) in comparison with the digested samples with added ingredients (oregano (1.82±0.10 g/100 g d.w) or beer (1.86±0.16 g/100 g d.w)), suggesting that oregano/beer could either promote a better digestion of fatty acids or prevent further LipOx during digestion (Table III.1). The formation of LOPs could be related with fatty acid oxidation, since the digested samples with added beer showed the lowest content of ω-3 and ω-6 PUFAs (Figure III.7), followed by digested samples without ingredients; and finally, the samples with oregano showing the minor losses of fatty acids. These data support all the results observed concerning MDA, HNE and HEX formation. Hur and colleagues (343) studied the impact of several cooking methods (oven cooking, grilling, boiling, and microwaving) on lipolysis and LipOx of pork patties during in vitro digestion reporting increased MDA values and cholesterol oxidation products, regardless of the cooking method. Moreover, Van Hecke and colleagues (335) recently reported a positive Pearson correlation between LOPs formation and PUFA content found in digested meat, with ω-3 PUFAs strongly correlating with TBARS values and HHE, whereas no correlation was found between ω-6 PUFAs and HNE.

CHAPTER III. RESULTS AND DISCUSSION

Figure III.7 – The variation of ω-3 and ω-6 fatty acids content in in vitro digested samples (oven cooked samples) using the raw samples as starting point content (raw content of each condition was subtracted to in vitro digested samples) to evaluate their stability to cooking. Values are means ± SD of three independent in vitro digestion experiment and each sample extracted twice (n=6). The statistical analysis was performed within digested samples with different letters meaning significant differences within each group (p<0.05).

The increased values of HEX and HNE found in samples cooked with oregano may be due to the loss of antioxidant ability of the herb against the oxidant agents present/formed during the digestive process. Indeed, Raes and colleagues (344) demonstrated that the in vitro digestion conditions (e.g. addition of saliva and acidic pH) influenced the antioxidant stability and activity of carnosic acid, quercetin and α-tocopherol. Van Hecke and colleagues (172) monitored the formation of MDA, HNE, and HEX after in vitro digestion of high-fat beef products with/without the addition of herbs/spices, among them oregano. Oregano exerted a clear antioxidant activity under digestive conditions when added at 0.5% or 1% doses before thermal treatment of meat. The same authors (290) also evaluated the antioxidant ability of several hydrophilic and hydrophobic phenolic compounds on LipOx while digesting of low- and high-fat beef samples, reporting pro-oxidant and antioxidant behaviours depending on their dose, with hydrophilic antioxidants (ascorbic, gallic, ferulic and chlorogenic acids) exhibiting a stronger preventive effect on low-fat beef samples. Moreover, adding α-tocopherol decreased LipOx during ex vivo gastric digestion of cooked turkey meat (345).

CHAPTER III. RESULTS AND DISCUSSION

On the other hand, the decreased content of aldehydes may be due to their degradation into tertiary oxidation products (e.g. volatile compounds) or interaction with protein residues. Indeed, some reports propose the binding of HNE with amino groups of protein in the intestinal phase, hindering its measurement in the free form (170). Moreover, the interactions between lipid-derived aldehydes and nitrogenated compounds can also take place through Maillard reactions, and esterification reactions between acids and alcohols of low molecular weight (318). Figure III.6 shows a massive increase in carbonylation after in vitro digestion, regardless of the addition of ingredients, keeping the same behaviour after cooking: samples with beer exhibited the highest carbonyls values. The increased carbonyls values after digestion are in agreement with Hu and colleagues (318). Likewise, a considerable increase on the fluorescent intensity was observed after in vitro digestion of samples (Figure III.8), at lesser extent in samples with added oregano, with the formation of only one peak and shift to ~460 nm, probably due to reaction of amino acids released during digestion with the reactive aldehydes formed during LipOx.

Figure III.8 – Variation of fluorescence in in vitro digested samples using the fluorescence of raw samples as starting point content (fluorescence of raw samples of each condition was subtracted to cooked samples) in order to investigate the formation of SB during digestion with and without ingredients. Values are means of three independent cooking/in vitro digestion experiments and each sample extracted twice (n=6).

Veberg and colleagues (342) verified that reactions between amino acids (e.g glycine and Lys) and unsaturated aldehydes (e.g MDA and 2-hexenal) exhibited a maximum fluorescence emission around 450-470 nm (excitation at 386 nm). In addition, a positive Pearson correlation has been reported between protein carbonylation and TBARS levels (335). The formation of only one peak had already been reported elsewhere (102, 318) and justified with extreme oxidation and consequent overlap of peaks. Thus, the high carbonyl and SB formation during in vitro digestion suggest a high ProtOx at the intestinal phase.

CHAPTER III. RESULTS AND DISCUSSION

Moreover, the contents of free Thr, Phe, and His decreased after digestion along with the increase of cystine (dimer of cysteine) (Figure III.9). The later one can be used as a measure of oxidation since the loss of free thiol groups have been used as protein marker of oxidation during storage, after cooking, and in vitro digestion (318).

Figure III.9– The variation of free amino acids content in in vitro digested samples (oven cooked samples) using the raw samples as starting point content (raw content of each condition was subtracted to digested samples) in order to investigate the stability of amino acids considered as susceptible to oxidation: (A) Polar, non-charged amino acids; (B) Aromatic amino acids, (C) Apolar amino acids, and (D) Polar, positively charged (90). The red lines correspond to the value found in oven cooked samples (before digestion). Values are means ± SD of three independent cooking/in vitro digestion experiments, each sample injected three times (n=9). The statistical analysis was performed within oven cooked or digested samples with different letters meaning significant differences within each group (p<0.05).

After digestion of a protein-rich product it was expected to observe an increase of free amino acids content as results of proteolysis. Met, Tyr, Trp, Val, Leu, Ile, and Lys contents increased in digested samples in comparison with raw and cooked samples, but for some of them the balance between digestion/oxidation was low (Figure III.9). Protein crosslinks, aggregation and polymerization as well as complexation with aldehydes may alter the protease-active sites, hindering the enzymatic proteolysis, and consequently reducing the 88

CHAPTER III. RESULTS AND DISCUSSION amino acid release during digestion (86, 318). Moreover, prolonged cooking of beef (100 ºC, 30 min) showed to impact the digestibility of peptides with MW < 25 kDa, unabling their broken down into individual amino acids and affecting the bioavailability of amino acids (346). Pepsin digestion is more prone to produce peptides containing Tyr, Phe, and Leu at the N-terminus or Met, Phe, and Leu at the C-terminus, while trypsin cleavages bonds in C- terminus of arginine and Lys (347). The abovementioned amino acids involved in the proteolytic reactions are susceptible to oxidation, which could limit the proteases activity on meat (346). In this sense, although cooking resulted in a moderate ProtOx, a remarkable increase of carbonyls and SB structures occurred during digestion. Nonetheless, despite the increase of LOPs, carbonyls and SB structures in samples cooked with oregano, this was the culinary practice with the lowest formation of oxidation markers after in vitro digestion and highest prevention of PUFAs. In this sense, oregano herbs should be added to meat prior to cooking to reduce the formation of hazardous compounds during cooking and digestion.

III.1.3 FINAL REMARKS

Cooking remarkably increased MDA, HEX, HNE, and SB structures, while carbonyls content remained equal. The oxidation observed was confirmed by the loss of free amino acids susceptible to oxidation and ω-6 PUFAs. Adding oregano decreased MDA, HEX, and HNE formation (p<0.05) and SB, as well as preserved ω-6 PUFAs content, while adding beer did not influence the formation of LOPs and contributed to the loss of PUFAs and some free amino acids. The in vitro digestion increased MDA, carbonyls, and SB structures regardless of the culinary practices, while HNE and HEX contents were only increased in samples cooked with oregano. The low free amino acids content observed after digestion suggest either impairment of enzymatic proteolysis hindering the release of free amino acid or their conversion into carbonyls and/or participation in SB formation. PUFAs also showed to be unstable during digestion, mostly in samples cooked without ingredients or with added beer. The addition of oregano prior to cooking contributed with the lowest formation of HNE, HEX, MDA, carbonyls, and Schiff bases either after cooking or in vitro digestion, as well as prevented the loss of PUFAs in both stages. Thus, this herb could be used as a mitigation strategy to reduce LipOx and ProtOx during cooking, preserve the nutritional quality of meat, as well as reduce the exposure of humans to LOPs after ingestion of cooked meat.

CHAPTER III. RESULTS AND DISCUSSION

III.2 ANTIBACTERIAL AND COCCIDIOSTAT DRUGS

Antibacterial drugs as β-lactams, sulfonamides (SAs), tetracyclines (TCs), macrolides, fluoroquinolones (FQs) and coccidiostats are widely used in poultry livestock production to prevent and treat diseases (348). Their administration via feed often leads to their presence as residues in animal tissues and consequent carry-over to humans causing allergic reactions in hypersensitive individuals, destruction of the dynamic balance of gastrointestinal flora, and increase of microbial resistance to antibiotics. As chicken meat is cooked before consumption it is important to study the effect of the different cooking practices on antibacterial and coccidiostat drugs (ACD) residues stability, a subject scarcely reported in bibliography. Moreover, no information was found regarding ACDs behaviour in the presence of culinary ingredients (e.g herbs and beer/wine), which have shown to reduce humans’ exposure to other chemical hazards in cooked meat. In this sense, this study aimed to investigate the effect of eight different culinary practices (oven or microwave cooking combined with the addition of herbs and/or beer) on the stability of 14 ACDs, representative of the abovementioned ACDs classes, and their bioaccessibility after in vitro digestion. This study evaluated for the first time the fate of amoxicillin, trimethoprim, and narasin during cooking and digestion. Depending on ACDs level of change, the presence of transformation products derived from cooking or in vitro digestion procedure was also investigated.

III.2.1 METHOD DEVELOPMENT AND VALIDATION

The linearity of the chromatographic method was determined by preparing matrix- matched calibration curves using 5 calibration points at concentration ranging from 10 to 1500 µg/kg in meat (raw and cooked) and 75 to 1500 µg/L in bioaccessible digested fractions (QuEChERS method), as recommended by SANTE guidelines since most of the compounds showed significant matrix effect (>20% or <-20%) (327). In the dilute and shoot method, the calibration curves were performed in methanol. Good linear response was obtained for all ACDs, with coefficient of determination (R2) > 0.997 for raw and cooked matrices. In the case of digested samples, a good linear response was observed (R2 > 0.991) except for CTC (R2=0.90). Table III.2 summarizes the recoveries of ACDs in spiked levels of 10 and 100 µg/kg in raw chicken breast that were higher than 70%, with relative standard deviations (%RSD) values lower than 20%, as well as the recoveries found in QuEChERS method for digested samples spiked at 100 and 200 µg/kg.

CHAPTER III. RESULTS AND DISCUSSION

75 75 75 75 75

LOQ

(µg/Kg)

(%RSD)

9.8 0.4 0.7 2.4 7.3 1.1 2.7 2.1

18.9 10.3

day day

Intra

day. %RSD) day.

99.7(6.5) 94.6(4.6) 93.8(3.7)

100.2(9.5) 109.0(7.4) 104.9(8.0) 113.0(4.1) 116.9(5.7) 97.1(21.4) 101.6(5.5)

Recovery (%) Recovery

DIGESTED FRACTION DIGESTED

(Inter

100 100 200 100 200 100 200 200 100 200

Spiked level Spiked(µg/Kg) level

10 10 10 10 10

LOQ

(µg/Kg)

day

8 5 4

4.8 4.2 5.7 3.4 1.8 8.4

12.1

(%RSD)

Intra

day. %RSD) day.

97.1(4.5)

92.7(12.0) 104.6(2.1) 98.8(14.6) 118.4(8.7) 107.6(6.3) 112.1(1.1) 101.2(7.0) 70.1(14.1) 89.4(10.3)

Recovery (%) Recovery

(Inter

RAW CHICKEN BREAST CHICKEN RAW

10 10 10 10 10

100 100 100 100 100

(continued)

–

Spiked level Spiked(µg/Kg) level

TYL

SDZ

TMP

SMX

ACDs

TableIII.2 SDMX

CHAPTER III. RESULTS AND DISCUSSION

On the other hand, in the dilute and shoot method, AMX and the four TCs showed very low recoveries (Table III.3). The low recoveries (<60%) found in the dilute and shoot method, where no extraction methodology was applied (i.e. only the matrix has an effect on ACDs), suggests that these ACDs are either unstable in the bioaccessible matrix (pH, electrolytes concentration) or interact with compounds of the matrix. Moreover, as the digestion process requires the addition of a huge amount of salts, including bivalent cations (Ca2+ or Mg2+), an excess of EDTA was added to free all TCs that could quelate with bivalent cations.

Table III.3 – Percentage of recovery and inter-day precision (%RSD) of ACDs for both dilute and shoot method used in bioaccessible samples (n=3), and the method used in non-bioaccessible fraction (NBIO, n=5). The mass balances percentages (%MB) of the in vitro digestion experiment are also presented.

Sample dilution NBIO

100 µg / L 500 µg / L MB (%)*

Recovery (%) RSD (%) Recovery (%) RSD (%)

AMX 21.3 24.1 - - -

CTC 33.9 9.2 100.6 0.6 52.4±21.8

DC 49.9 19.1 117 0.4 57.8±37.5

OTC 57.5 16 112 0.4 83.9±42.0

TC 39.2 21.2 117 0.7 65.0±18.5

ENR 84.5 2.5 109 0.5 55.0±22.0

CIP 85.9 3.6 112 1.6 51.1±19.7

NAR 85.7 8.5 102 1.2 42.0±17.2

DNC 84.1 8.9 113 0.2 -

SDZ 89.2 1.9 111 1.8 73.8±28.3

SDMX 86.7 4.5 112 0.8 89.7±24.7

SMX 89.7 2.2 114 2.7 110.9±5.29

TMP 87.6 2.4 114 0.9 75.9±45.8

TYL 79.5 2.4 128 3.7 69.7±45.6

*MB was calculated as MB = ((ng ACD in bioaccessible fraction + ng ACD in non bioaccessible fraction)/ ng ACD in cooked samples before digestion) x 100

All other ACDs presented recoveries higher than 79.5%, with %RSD lower than 20%. Concerning the matrix effects observed in the dilute and shoot method and the high dilution factor of the method (8x in digestion + 5x with MeOH) that would preclude the quantification of some compounds, it was decided to use the QuEChERS method to quantify the ACDs in 93

CHAPTER III. RESULTS AND DISCUSSION

digested matrices. The LOQs were set at 10 µg/kg in chicken muscle for all ACDs except AMX (25 µg/kg) and 75 µg/L in the bioaccessible fraction for all ACDs.

III.2.2 IMPACT OF CULINARY PRACTICES

The impact of the 8 cooking practices – oven or microwave with/without the addition of herb and/or beer – on 14 ACDs is shown in Figures III.10 until Figure III.17. The contents of CTC, TC, SDZ, SMX, TMP and TYL decreased using oven and microwave-based methods, regardless of the initial level of fortification. Oven-based methods also decreased AMX contents. Considering the amount of ACDs studied as well as the number of culinary practices evaluated, the discussion of results will be presented separately in sub-sections, by class of ACD, for a clearer interpretation.

III.2.2.1 AMOXICILLIN

The reduction of AMX content was higher in oven samples (35 to 72%), regardless the initial level of fortification (Figure III.10A), than in microwaved samples with maximum reduction of 28% (Figure III.10B).

Figure III.10 - The impact of oven cooking (A) and microwaving (B) on amoxicillin. Data is expressed as the percentage (%) of ACD that remained after cooking with the addition or not of herbs and/or beer using two initial levels of fortification: 0.1 mg/kg and 1 mg/kg. The dashed line relates to the initial concentration quantified in raw samples, normalized to 100%; Ov – Oven cooking; OvH – oven cooking with herbs; OvB – oven cooking with beer; OvHB – oven cooking with herbs and beer; Mw – Microwaving; MwH – Microwaving with herbs; MwB – Microwaving with beer; MwHB – Microwaving with herbs and beer. Columns with different letters in each ACD concentration differ significantly (p<0.05).

CHAPTER III. RESULTS AND DISCUSSION

The samples fortified at 0.1 mg/kg significantly differed (p<0.05) between Ov and OvH (decrease) or Mw and MwB (increase), suggesting some influence of adding herbs or beer on AMX stability. On the other hand, samples fortified at the highest level (1 mg/kg) did not differ among culinary practices. Thus, the initial level of fortification seemed to influence the reduction of AMX during oven cooking. To the best of our knowledge there are no reports concerning the AMX stability on meat during cooking. Hsieh and colleagues (147) studied the stability of AMX in water at two distinct concentrations (50 mg/mL and 200 mg/mL) under high temperature (100 ºC and 121 ºC over 15 min) reporting this drug as thermal labile with reducing percentages of 40 and 80% at 100 ºC and 121 ºC, respectively, regardless the concentration tested. Contrary to our results, the initial amount of antibiotic did not interfere with AMX reduction. Concerning the high reduction of AMX after oven cooking it would be expected to find some transformation products (TPs). No reliable TPs were found using Compound Discover® and MassChemSite®, however, high-resolution mass spectrometry (HRMS) analyses shown a new peak formation (RT=7.31-7.41 min) on oven cooked meat fortified with AMX when compared with fortified raw meat, as shown in Figure III.11.

CHAPTER III. RESULTS AND DISCUSSION

Figure III.11 - Typical ion chromatograms (m/z= 366.11209) of AMX (RT=5.46 min) from HPLC-Orbitrap-MS analysis made in raw, cooked, and digested samples showing the isomerization of AMX into another product sharing the same m/z but different retention time (RT=7.31 min).

The rearrangement of AMX into two different molecules has been reported by Deschamps and colleagues (349). Hsieh and colleagues (147) also observed the formation of new peaks on unknown identity along AMX degradation. Also, a 2 to 8-fold increase of microbial inhibition concentration (MIC) suggested a loss on antibacterial resistance efficiency. β-lactam low stability under heating is mainly due to the hydrolysis susceptibility of the high ring strain of the small β-lactone ring (139).

CHAPTER III. RESULTS AND DISCUSSION

III.2.2.2 FLUOROQUINOLONES

FQs were quite stable to cooking with reductions lower than 38% (Figures III.12A and Figure 12B).

Figure III.12 - The impact of oven cooking (A) and microwaving (B) on fluoroquinolones. Data is expressed as the percentage (%) of ACD that remained after cooking with the addition or not of herbs and/or beer using two initial levels of fortification: 0.1 mg/kg and 1 mg/kg. CIP – ciprofloxacin, ENR – enrofloxacin. The dashed line relates to the initial concentration quantified in raw samples, normalized to 100%; Ov – Oven cooking; OvH – oven cooking with herbs; OvB – oven cooking with beer; OvHB – oven cooking with herbs and beer; Mw – Microwaving; MwH – Microwaving with herbs; MwB – Microwaving with beer; MwHB – Microwaving with herbs and beer. Columns with different letters in each ACD concentration differ significantly (p<0.05).

The high stability of FQs herein described is in agreement with Lolo and colleagues (152) and Roca and colleagues (350). Roca and colleagues (350) reported a maximum loss of 12% of CIP after treating milk at 120 ºC for 20 min. ENR and CIP content also differed concerning the initial level of fortification, mostly for ENR in samples fortified at 0.1 mg/kg. ENR content was decreased (p<0.05) during microwaving (all culinary practices) and oven cooking with the addition of extra ingredients. Precisely, a decrease in ENR content was observed during oven cooking with OvHB having the lowest amount (62% after cooking) (p<0.05). This may suggest a synergistic effect between herbs and beer on ENR reduction. This effect was not observed in microwaved samples, with Mw being the most efficient method (p<0.05) and OvB the least efficient in compounds reduction. In the samples fortified at 0.1 mg/kg, the decrease of CIP content was 16% (p<0.05) in OvHB (0.1 mg/kg level) samples, whereas in Mw and MwH samples, the decrease was 12% and 5%, respectively. The different behaviour observed between oven and microwaving may result from the different oxidative stages achieved depending on the cooking method used and the fact that FQs may be degraded by oxidative processes (351, 352). Moreover, the presence of phenolic antioxidants have shown to inhibit the oxidation of antibiotics avoiding their degradation upon oxidative processes (353). Thus, the higher ENR and CIP contents 97

CHAPTER III. RESULTS AND DISCUSSION

observed in cooked samples in the presence of herbs and/or beer in comparison with microwaving/oven cooking could be related to that (antioxidants protecting FQs from degradation). Lolo and colleagues (152) reported the influence of several cooking procedures (Mw, roasting, boiling, grilling, and frying) on naturally incurred chicken muscle, concluding that ENR was also stable during cooking. They verified an apparent decrease of ENR in microwaved samples, and an apparent increase in roasted samples, and justified the first phenomenon with the migration of ENR into surrounding liquid/meat juices, and the second with the lower moisture of the cooked samples. However, none of these hypotheses justifies the decrease of FQs herein observed because if migration into surroundings happened, the FQs decrease would be observed in all samples, which did not happen, and concerning the moisture content, corrections were made in cooked samples to allow comparison with the raw controls. In this sense, the reductions observed for CIP and ENR in Ov and Mw samples result from their thermal degradation.

III.2.2.3 TETRACYCLINES

TCs are among the most unstable antibacterial drugs described in literature (354). Figures III.13A and Figure 13B demonstrate that TCs stability to cooking is not structure- related with the 4 TCs showing distinct behaviours toward cooking even though sharing similar structures. Moreover, the initial level of fortification also influences the reduction of CTC and OTC during cooking. All cooking methods reduced at least 9.6% of TC (Ov) and 12.6% of CTC (MwHB), regardless the initial fortification level, while OTC behaviour depended on fortification level, and DC was quite stable during all cooking methods. In this sense, TCs could be ranked as DC (cooking-stable) > TC = OTC > CTC (cooking-labile). DC content was only decreased after Ov and OvB with a maximum reduction of 13%, regardless the level of fortification, while Mw, MwH, and MwB reduced DC contents only on 1 mg/kg fortified samples, with a maximum reduction of 35% observed in Mw. DC was also reported as the most cooking stable TCs after microwaving (2450 MHz, 6 power level, 10- 20 min) and roasting (180 ºC, 40-80 min) chicken breast, with reductions of 18.2 and 13.7% after Mw and roasting at lesser time (10 and 40 min) (140). Our findings are in agreement with the previous research since similar reductions were found using smaller times of cooking.

CHAPTER III. RESULTS AND DISCUSSION

Figure III.13 - The impact of oven cooking (A) and microwaving (B) on tetracyclines. Data is expressed as the percentage (%) of ACD that remained after cooking with the addition or not of herbs and/or beer using two initial levels of fortification: 0.1 mg/kg and 1 mg/kg. CTC – chlortetracycline, DC – doxycycline; OTC – oxytetracycline; TC – tetracycline The dashed line relates to the initial concentration quantified in raw samples, normalized to 100%; Ov – Oven cooking; OvH – oven cooking with herbs; OvB – oven cooking with beer; OvHB – oven cooking with herbs and beer; Mw – Microwaving; MwH – Microwaving with herbs; MwB – Microwaving with beer; MwHB – Microwaving with herbs and beer. Columns with different letters in each ACD concentration differ significantly (p<0.05).

TC was stable during cooking with reductions ranging from 18 to 38% in microwaved samples and from 9 to 32% in oven cooked samples. Similar behaviours were observed in all microwaved samples, while in the case of oven cooked samples opposite responses were observed depending on level of fortification. Higher reductions for TC than those found here were reported: Abou-raya and colleagues (140) reported TC reduction of 40.6 and 46% after microwaving (10 min) and roasting (180 ºC, 40 min) chicken breast, respectively; Gratacos-Cubarsi, Fernandez-Garcia (144) observed reductions ranging 85-90% after microwaving (440 W, 45s) and boiling (100 ºC, 2-14 min) chicken breasts, as well as the formation of anhydrotetracycline (ATC) and 4-epitetracycline (4eTC). Considering the low reductions on TC content after cooking observed in our experiments, it was not expected to 99

CHAPTER III. RESULTS AND DISCUSSION

find any of these TPs from our HRMS analysis, and actually, no TPs were detected by HRMS. OTC contents were slightly decreased after oven cooking samples at the highest fortification level (1 mg/kg) with Ov showing the maximum reductions, and no decreased contents in the microwaved ones. At the lowest level of fortification, OTC content was considerably lower after cooking samples, and reductions ranged from 27 to 38% in microwaved samples and 50% in oven cooked ones. Higher reductions of OTC in Mw and Ov samples suggested a negative effect of adding herbs/beer. Contrary to our findings, OTC has been reported as a highly degradable ACD with reductions ranging 25-99% under several thermal treatments (microwave, boiling, roasting, grilling, and frying) in meat (140, 146, 355-357). Moreover, migration of OTC into surrounding liquid or meat juices was observed by some researchers, which may explain the high loss of OTC, since some reports did not describe the formation of TPs as 4-eOTC, α-apo-OTC and β-apo-OTC. Considering the high stability of OTC in fortified samples at 1 mg/kg it was not expected to find any of this TP in our HRMS analysis. CTC was the most cooking labile TCs herein studied with reductions higher than 50%, except Mw at the lowest fortification level. The addition of herbs/beer before oven cooking did not significantly alter CTC content in samples at the lowest level of fortification; whereas an increasing tendency was observed at the highest level of fortification with OvHB exhibiting the highest content, and the same behaviour was observed in microwaved samples. Only one report was found concerning CTC stability during cooking of meat reporting lower reduction after microwaving (30%) and roasting (32%) chicken breasts (140). The differences may be due to different levels of contamination, different meat matrices and internal temperatures achieved since they cooked an entire portion of chicken breast, while in our experiments we used minced meat, which could help reduce CTC content. Hsieh and colleagues (147) studied the stability of CTC in water at 121 ºC over 15 min reporting degradations around 60%. CTC contents were highly reduced, thus it was expected to find some TPs. The MassChemSite® software was able to identify one TP derived from the dechlorination of CTC. Also, analysing the ion chromatogram obtained in raw and cooked samples (controls and fortified), a decrease of CTC (RT=7.75 min) was observed along with the increase of other peaks (RT=6.80-7.40 min) sharing the same m/z as CTC (Figure III.14).

100

CHAPTER III. RESULTS AND DISCUSSION

Figure III.14 - Typical ion chromatograms (m/z= 479.12201) of CTC (RT=7.75 min) from HPLC-Orbitrap-MS analysis made in raw, cooked, and digested samples showing the isomerization/epimerization of CTC into another products sharing the same m/z but different retention time (RT=6.80-7.40 min).

Loftin and colleagues (143) pointed out the epimerization of the dimethylamino group attached to the position 4 of TCs as the initial degradation product of CTC and TC. Moreover, CTC molecule rearrangement through isomerization/epimerization into 4-epi- chlortetracycline (4-epi-CTC), isochlortetracycline (iso-CTC) and 4-epi-iso-chlortetracycline (4-epi-iso-CTC) has been reported (143), keeping the same m/z than the parent compound. Thus, these peaks may confirm the loss of CTC through isomerization/epimerization reactions. OTC and TC would also suffer isomerization and epimerization (143), however,

101

CHAPTER III. RESULTS AND DISCUSSION

due to the lower reductions observed after cooking in samples fortified at 1 mg/kg it was not possible to find any of these compounds or other transformation products.

III.2.2.4 SULFONAMIDES

In general, almost all cooking methods reduced SDZ, SDMX, SMX and TMP contents, with SMX in samples fortified at 0.1 mg/kg showing reductions higher than 50% (Figure III.15A and Figure III.15B).

Figure III.15 - The impact of oven cooking (A) and microwaving (B) on sulfonamides. Data is expressed as the percentage (%) of ACD that remained after cooking with the addition or not of herbs and/or beer using two initial levels of fortification: 0.1 mg/kg and 1 mg/kg. SDZ – sulfadiazine; SDMX – sulfadimethoxine; SMX – sulfamethoxazole; TMP – trimethroprim. The dashed line relates to the initial concentration quantified in raw samples, normalized to 100%; Ov – Oven cooking; OvH – oven cooking with herbs; OvB – oven cooking with beer; OvHB – oven cooking with herbs and beer; Mw – Microwaving; MwH – Microwaving with herbs; MwB – Microwaving with beer; MwHB – Microwaving with herbs and beer. Columns with different letters in each ACD concentration differ significantly (p<0.05).

102

CHAPTER III. RESULTS AND DISCUSSION

SAs behaviour was similar regarding the initial level of fortification. SDMX was the most cooking-stable SA with maximum reductions of 28% and differing with the level of fortification and type of cooking. The influence of herb/beer was only significant in samples fortified at 1 mg/kg with lower reductions observed in OvH, OvHB, MwB, MwHB. This suggests a negative effect on adding herbs/beer. SDZ reductions ranged from 22-42% and 11-42% in oven and microwaved samples, respectively. SMX was the most cooking-labile SA, with reductions ranging 32-66% in cooked samples fortified at 0.1 mg/kg. The samples fortified at 1 mg/kg presented maximum reduction of 41%. This suggests that SMX stability may be concentration dependent. Even though the reductions were considerable no TPs were identified. Ismail-Fitry and colleagues (358) reported reductions of about 60% for SDZ and SMX after cooking chicken meatballs (boiling, deep-frying and microwaving). Furusawa and colleagues (359) studied similar cooking conditions – microwaving (500 W, 1 min) and roasting (170 ºC, 12 min) reporting reductions of 36% for SDZ and SMZ in microwaved samples and 4 and 39% for SDZ and SMZ after roasting, respectively. Adding herbs before microwaving significantly reduced SMX content (p<0.05), regardless the level of fortification, while in oven cooked samples this effect was only observed in the fortified samples at the lowest level (OvH). In the case of SDZ higher reductions were observed when adding herbs prior to cooking, proposing a beneficial influence in decreasing SDZ. The chicken meatballs studied by Ismail-Fitry and colleagues (358) also contained condiments (corn starch, black pepper, sugar and garlic), and since our results show a higher reduction on both SDZ and SMX with the addition of extra- ingredients, this could explain the higher reduction observed in microwaved chicken meatballs. Although, the presence of phenolic antioxidants during photodegradation inhibits the oxidation of sulfonamides (SDZ and SMX included) in aqueous solutions by reducing the SAs-intermediate oxidation products back to the parent compound (353). The complexity of the matrices studied (meat and oregano) that contain both antioxidant and pro-oxidant compounds, may justify the reducing effects observed on SDZ and SMX with the addition of herbs. TMP reduction was similar in all cooking methods with reductions ranging from 21 to 54%, regardless of the level of fortification. The addition of extra ingredients was only significant (p<0.05) in microwaved samples with herbs. Although no literature reports were found concerning TMP reduction in cooked chicken, 23% reduction of this ACD was observed in cooked bivalves (360), which corroborates the low reduction of TMP.

103

CHAPTER III. RESULTS AND DISCUSSION

III.2.2.5 MACROLIDES

TYL can be pointed out as the most cooking-labile ACD herein studied, as all cooking practices reduced its content from 32% up to 77% (Figures III.16A and Figure III.16B).

Figure III.16 - The impact of oven cooking (A) and microwaving (B) on tylosin. Data is expressed as the percentage (%) of ACD that remained after cooking with the addition or not of herbs and/or beer using two initial levels of fortification: 0.1 mg/kg and 1 mg/kg. The dashed line relates to the initial concentration quantified in raw samples, normalized to 100%; Ov – Oven cooking; OvH – oven cooking with herbs; OvB – oven cooking with beer; OvHB – oven cooking with herbs and beer; Mw – Microwaving; MwH – Microwaving with herbs; MwB – Microwaving with beer; MwHB – Microwaving with herbs and beer. Columns with different letters in each ACD concentration differ significantly (p<0.05).

Significant differences (p<0.05) were observed on TYL content after cooking Mw/Ov samples and samples with the addition of herbs and beer (MwHB/OvHB), pointing that adding ingredients does not benefit TYL reduction. Although samples fortified at 1 mg/kg were highly decreased no TP were identified. The influence of initial content on TYL reduction was reported by Salaramoli and colleagues (150). These authors studied the impact of boiling and microwaving on chicken meats, reporting lower reductions (14-21%) in microwaved chicken meatballs, however boiling for 10 min resulted in reductions higher than 60%. This suggests that some loss of TYL may be due to migration into surrounding liquids considering that no TPs were found to justify the loss of TYL during cooking.

III.2.2.6 COCCIODIOSTATS

NAR and DNC highest reductions were observed in Ov and Mw samples while the addition of herb/beer negatively influenced their reductions (Figure III.17A and Figure III.17B). Higher reductions of NAR were observed ranging from 23 to 45% (microwaved) and 23 to 62% (oven cooked). The initial level of fortification seems also to influence NAR

104

CHAPTER III. RESULTS AND DISCUSSION degradation during oven cooking. To the best of our knowledge this is the first study regarding the influence of cooking on NAR content in chicken meat. From our findings it is possible to conclude that NAR is unstable during cooking, its reduction is concentration dependent being less stable at more realistic concentration (0.1 mg/kg), and the addition of herbs/beer does not promote its reduction.

Figure III.17 - The impact of oven cooking (A) and microwaving (B) on coccidiostats. Data is expressed as the percentage (%) of ACD that remained after cooking with the addition or not of herbs and/or beer using two initial levels of fortification: 0.1 mg/kg and 1 mg/kg. NAR – Narasin, DNC - 4,4’-dinitrocarbanilide. The dashed line relates to the initial concentration quantified in raw samples, normalized to 100%; Ov – Oven cooking; OvH – oven cooking with herbs; OvB – oven cooking with beer; OvHB – oven cooking with herbs and beer; Mw – Microwaving; MwH – Microwaving with herbs; MwB – Microwaving with beer; MwHB – Microwaving with herbs and beer. Columns with different letters in each ACD concentration differ significantly (p<0.05).

In the case of DNC, a similar decrease was observed after microwaving and oven cooking, with maximum reductions of 29% and 28% after microwaving and oven cooking, respectively. Contrarily to our findings, degradations higher than 50% after several cooking methods (boiling, grilling, microwaving, frying, and roasting) were reported in incurred chicken breast as well as the identification of a product from DNC hydrolysis (p-nitroaniline, p-NA) (249, 361). p-NA is becoming a compound of concern in cooked chicken tissue due to its toxicity and probable involvement with cancer and methemoglobinemia diseases (361). Thus, DNC presence in meat becomes a dual problem: cooking with herbs does not degrade DNC and it will consequently be present in cooked meat, while cooking without herbs leads to DNC degradation and formation of a potential harmful compound (p-NA) that will appear in cooked meat. From our findings it seems that adding herbs prevents human exposure to p-NA, although, it does not help reduce DNC exposure to humans. Thus, new strategies should be studied to mitigate DNC in chicken meat avoiding the formation of concerned compounds like p-NA.

105

CHAPTER III. RESULTS AND DISCUSSION

III.2.3 ANTIBACTERIAL AND COCCIDIOSTAT DRUGS BEHAVIOUR TOWARD IN VITRO

DIGESTION

The influence of oven cooking and microwaving on the bioaccessibility of 14 ACDs is showed in Tables III.4 and Table III.5, respectively. The in vitro digestion experiments were carried out in both samples fortified at 0.1 mg/kg and 1 mg/kg, however, it was not possible to quantify the ACDs in samples fortified at the lowest level. In this sense, the bioaccessibility values shown in the following tables concern the digestion of cooked samples fortified at 1 mg/kg.

Table III.4 - Percentage (%) of bioaccessibility of ACDs in oven cooked samples (n=6) after duodenal in vitro digestion. Bioaccessibility (%) ACD Ov OvH OvB OvHB AMX

Overall, all ACDs presented bioaccessibilities lower than 60%, with SAs exhibiting the highest values. AMX and DNC were not detected in any sample after in vitro digestion and CTC was only detected in microwaved samples with bioaccessibility percentages ranging from 15 to 18.5%. DC and TYL showed the lowest bioaccessibility (<20%). The addition of extra ingredients did not significantly (p<0.05) change the bioaccessibility of CTC, DC and TYL. OTC and TC presented similar bioaccessibility ranging from 20.5 to 44.9%, with cooking methods with the addition of herbs (OvH, OvHB, MwH, and MwHB) expressing the highest percentages of bioaccessibility.

106

CHAPTER III. RESULTS AND DISCUSSION

Table III.5 - Percentage (%) of bioaccessibility of ACDs in microwaved samples (n=6) after duodenal in vitro digestion. Bioaccessibility (%) ACD Mw MwH MwB MwHB AMX

An increased bioaccessibility on samples cooked with herbs (OvH and MwH) was also observed in SAs, except SDMX in OvH. SAs exhibited the highest percentages of bioaccessibility with SDMX being the least (21.9-40.4%) and SMX the more bioaccessible (41.0-61.3%). The addition of extra ingredients altered the bioaccessibility of SAs, increasing SDZ and SMX in oven cooked samples, and MwH samples. TMP had lower bioaccessibility than SAs (16.1-23.9%) with the addition of herbs/beer significantly (p<0.05) increasing values in oven cooked samples. To the best of our knowledge no reports were found concerning the impact of cooking on ACD bioaccessibility. Nonetheless, the aqueous solubility, drug lipophilicity, co-ingestion foodstuffs, and pKa in relation to gastrointestinal profile have been pointed out as the key factors determining the absorption of a drug (183). The compounds need to be sufficiently water soluble to dissolve in gastric and intestinal fluids (184), and the ACDs herein studied have different water solubility. According to the coefficient of partition (LogP) values that measures the hydrophobicity of a molecule (i.e. the higher the value, the lower the water solubility of the compound), the TCs are the most hydrophilic compounds (more water soluble) herein studied having negative coefficient of partition (LogP), while TYL, NAR, and DNC are the most hydrophobic compounds ones (less water soluble) with LogP > 1.63 (Table AI.2). Thus, as the bioaccessible matrix is mostly composed by water and knowing the water solubility of the different ACDs it would be expected high bioaccessibility of TCs (more water soluble), intermediate bioaccessibility of SAs, FQs and AMX, and low 107

CHAPTER III. RESULTS AND DISCUSSION

bioaccessibility of NAR, DNC and TYL (less water soluble), which was not always verified. Also, the matrix may exert some effects on ACDs bioaccessibility. Indeed, drug-food interactions are complex and have been widely studied in the last decades (184, 299, 300, 362, 363), with food constituents influencing the availability of a drug either enhancing or decreasing it. Food may influence the drug absorption by direct adsorption of drug onto food components, by chelation with polyvalent metal ions (Ca2+, Mg2+, Fe2+), or by complexation with proteins (184). Bushra and colleagues (299) reviewed the impact of several food constituents on antibiotics bioavailability reporting different bioavailabilities concerning the ingestion of high-fat meals, vegetables, milk, and charbroiled meals. The low mass balance percentages %MB = ((ng ACD bioaccessible + ng non-bioaccessible) / ng in cooked sample) x 100 (Table III.3) may suggest that some ACDs can still be present in the bioaccessible fraction, although not in its free form to be detected by the LC-QqQ-MS/MS analysis. Even though EDTA was added during the extraction, the presence of Ca2+ and Mg2+ derived from digestion experiments could justify the low bioaccessibility of free TCs since these drugs have a high affinity to form chelates with divalent cations resulting in insoluble or poorly absorbed complexes along the GIT (209). Additionally, TC was found to interact with proteins (364). Nielsen and Gyrd-Hansen (365) studied the bioavailability of OTC, TC, and CTC after oral administration to fed and fasted pigs, reporting low bioavailabilities for all TCs and only TC availability differed in fasted or fed pigs. Hörter and Dressman (183) also reviewed that drug solubility can be enhanced by amphiphilic bile components such as bile salts, lecithin and monooleins, resulting in their solubilisation into bile salts micelles increasing the solubility of poor soluble drugs. Indeed, Glanzer and colleagues (182) reported the interaction between macrolides and bile acids. Thus, the decreased bioaccessibility of almost all ACDs may be justified by the interaction of ACD with matrix components (cations, proteins or bile salts), hindering the quantification by LC-QqQ-MS/MS. This means that the low bioaccessibility observed may not give a realistic information of ACDs bioaccessibility because some might be bioaccessible while complexed with bile salts. Moreover, for those ACDs with acceptable %MB but low bioaccessibility (OTC, SDMX, and TMP), the remaining content of ACD in non-bioaccessible fraction will follow fermentation into colon interacting with bacteria and several studies show that human intestinal microbiome can be altered by the ingestion of antimicrobial drug residues in food derived from animals (366, 367). Additionally, a chronic exposure of humans to contaminated meat with ACDs may lead to a modulation of the intestinal bacterial community with potential health consequences (368).

108

CHAPTER III. RESULTS AND DISCUSSION

III.2.4 EXPLORATORY AND STATISTICAL ANALYSIS

As some variables (e.g. type of ACD, fortification level, and type of cooking) seem to determine the extent of ACD after cooking and in vitro digestion, an analysis of covariance

(ANCOVA) was performed to investigate the influence of each variable: ACD (X1), level of fortification (X2), type of cooking (X3), herbs (X4), and beer (X5) as well as their interactions on percentage of ACD after cooking and on their percentage of bioaccessibility (Table III.6). Concerning cooking data, a good model fit with R2 of 0.812 and R2 adjusted of 0.801, F- test value of 70.249 and root mean square errors (RSME) of 10.899 was built. The scatter plot of the correlation between predicted vs observed % of ACD after cooking is shown in Figure III.18. ACDs, fortification level and type of cooking were the main variables with statistical significance (p<0.05). The mean comparison of ACDs contents from all microwaved or oven cooked samples, regardless of other variables studied (type of ACD, concentration, herbs, beer) showed that oven cooked samples significantly have the lowest percentages of ACDs after cooking, suggesting oven cooking as a better method do reduce ACDs (Appendix II, Figure AII.1). The addition of herbs and/or beer were found not significant concerning their impact on ACDs % after cooking. Concerning the interactions, 8 of 10 interactions were of statistical significance: the level of fortification significantly interacts with all other variables, the ACDs interact with the type of cooking or herbs, and the type of cooking interacts with herbs or beer.

109

CHAPTER III. RESULTS AND DISCUSSION

Table III.6 - Summary of ANCOVA analysis for cooking and bioaccessibility regarding different cooking methods applied. Sum of Variables DF Mean squares F value p-value squares Cooking Model 75 6.26E+05 8.35E+03 7.02E+01 < 0.0001 Concentration (X1) 1 9.80E+03 9.80E+03 8.25E+01 < 0.0001 Drug (X2) 13 1.75E+05 1.35E+04 1.13E+02 < 0.0001 Type of cooking (X3) 1 1.97E+04 1.97E+04 1.66E+02 < 0.0001 Herbs (X4) 1 3.12E+02 3.12E+02 2.62E+00 ns Beer (X5) 1 2.82E+02 2.82E+02 2.37E+00 ns X1 x X2 13 1.23E+05 9.44E+03 7.95E+01 < 0.0001 X1 x X3 1 1.24E+04 1.24E+04 1.04E+02 < 0.0001 X1 x X4 1 3.30E+03 3.30E+03 2.78E+01 < 0.0001 X1 x X5 1 2.32E+03 2.32E+03 1.96E+01 < 0.0001 X2 x X3 13 6.21E+04 4.78E+03 4.02E+01 < 0.0001 X2 x X4 13 2.78E+04 2.14E+03 1.80E+01 < 0.0001 X2 x X5 13 2.47E+03 1.90E+02 1.60E+00 ns X3 x X4 1 1.15E+03 1.15E+03 9.67E+00 0.002 X3 x X5 1 8.60E+02 8.60E+02 7.24E+00 0.007 X4 x X5 1 2.82E+02 2.82E+02 2.38E+00 ns Error 1.22E+03 1.44E+05 1.19E+02 Goodness of the fit R² 0.812 Adjusted R² 0.801 RMSE 1.09E+01

Bioaccessibility Model 58 1.78E+05 3.07E+03 8.42E+01 < 0.0001 Drug (X2) 13 1.57E+05 1.21E+04 3.31E+02 < 0.0001 Type of cooking (X3) 1 3.25E+01 3.25E+01 0.891 ns Herbs (X4) 1 8.01E+02 8.01E+02 2.20E+01 < 0.0001 Beer (X5) 1 9.86E+02 9.86E+02 2.71E+01 < 0.0001 X2 x X3 13 1.22E+04 9.35E+02 2.56E+01 < 0.0001 X2 x X4 13 5.67E+03 4.36E+02 1.20E+01 < 0.0001 X2 x X5 13 2.64E+03 2.03E+02 5.57E+00 < 0.0001 X3 x X4 1 88.013 88.013 2.42E+00 ns X3 x X5 1 9.02E+02 9.02E+02 2.47E+01 < 0.0001 X4 x X5 1 27.09 27.09 0.743 ns Error 608 2.22E+04 3.65E+01 Goodness of the fit R² 0.889 Adjusted R² 0.879 RMSE 6.037 ns, not significant. Level of significance (p<0.05)

110

CHAPTER III. RESULTS AND DISCUSSION

ACDs stability during cooking does not rely only on one variable, but on the interaction of at least two variables. Moreover, the interaction ACD x cooking type shows that the stability of an ACD family cannot be predicted by assessing the thermal stability of only one compound of a certain family, but each compound must be tested individually. This behaviour is clearly observed in TCs family, which differed among TCs.

Figure III.18 - Analysis of covariance (ANCOVA). Scatter plot of the correlation models between predicted vs observed % ACD after cooking (A) and predicted vs observed % of ACD bioaccessibility (B).

In the case of ACDs bioaccessibility, the ANCOVA analysis resulted on a good model fit with R2 of 0.889 and R2 adjusted of 0.879, F-test value of 75 and root mean square errors (RSME) of 6.037, determining ACDs, herbs and beer addition as the main variables with statistical significance (p<0.05) (Table III.6 and Figure III.18). In this sense, ACDs bioaccessibility values are dependent on the type of drugs as well as are influenced by the addition of extra ingredients. Apart from the main variables, 4 of 6 interactions were considered of statistical significance: the ACDs significantly interact with all other variables, and the type of cooking interacts with beer. This shows that ACDs bioaccessibility does not rely only on one variable, but on the interaction of at least two variables.

III.2.5 FINAL REMARKS

Our findings helped to understand that most ACDs herein studied are stable during cooking exhibiting low reductions and therefore no TPs were identified, except for AMX, CTC and TYL that were highly reduced (>50%), with AMX and CTC reduction resulting from molecular rearrangement and CTC dechlorination reaction. This experiment was performed using fortified samples instead of naturally incurred chicken meat, and not all literature

111

CHAPTER III. RESULTS AND DISCUSSION

results using incurred meat agree with the present results, suggesting that some ACDs can exhibit different behaviours if naturally incurred. Adding herbs precludes the reduction of most ACDs, however, in the case of DNC, this phenomenon avoids the release of p-nitroaniline, a compound of concern in cooked chicken meat. Also, the amount of ACDs found after cooking is significantly influenced by the initial level of contamination, type of cooking, and ACD type, which means that the cooking stability of one class of ACD family cannot be predicted by studying only one family-related compound (e.g. TCs family). By means comparison, oven cooking seems to promote higher ACDs reductions than microwaving. Concerning the in vitro digestion, bioaccessibility up to 60% were observed for all ACDs however, for risk assessment purposes, it should be highlighted that ACD-bile salts interact, increasing the absorption of lipophilic compounds. Furthermore, drug-food interactions with cations, proteins and bile salts from matrix, hindered ACDs quantification by LC-QqQ- MS/MS, therefore their bioaccessibility prediction shouldn’t be exclusively based on the determination of its free form. In this sense, all data herein discussed elucidates some aspects concerning the influence of the level of contamination on ACDs degradation; their behaviour toward cooking practices, with/without the addition of herbs and/or beer, and the food-ACD interactions during in vitro digestion which introduce some instrumental limitations to assess final intake of these compounds.

112

CHAPTER III. RESULTS AND DISCUSSION

III.3 MYCOTOXINS

The use of livestock feed contaminated by mycotoxins contributes to animals’ exposure to toxic compounds and consequent carry over to humans. The use of roasting bags along with some seasonings to cook chicken has become a trend to keep the tenderness of meat, however, their impact on mycotoxins stability and in vitro bioaccessibility is currently unknown. In this sense, as a first approach, this thesis aimed to evaluate the impact of eight different cooking methods (oven or microwave with/without commercial herbs and/or roasting bags) on the stability of the 10 most prevalent mycotoxins (Chapter III.3.2) and their consequent in vitro bioaccessibility (Chapter III.3.3). This study comprised the first comprehensive outlook on mycotoxins fate during oven and microwaving for all 10 mycotoxins in meat, since knowledge on literature only existed for OTA after frying/boiling or AFB1/AFB2 after frying/broiling. As a second approach, focusing the mycotoxins interplay at the cellular level, this thesis aimed to study the absorption of mycotoxins through gastrointestinal epithelium (Chapter III.3.4) as well as their toxic effects in target organs, isolated and in combination (Chapter III.3.5), since the majority of absorption and toxicological data is obtained based on individual exposure. Moreover, we are concomitantly exposed to a mixture of toxic compounds in the environment (247), and mycotoxins are a good example of the importance of combined toxicity evaluation as they often co-occur in food (47, 268, 269). In addition, no information was found concerning gastric absorption of mycotoxins. In this sense, four prevalent mycotoxins from 4 distinct chemical groups/families – aflatoxins (AFB1), trichothecenes (DON), fumonisins (FB1), and ochratoxins (OTA) were selected to investigate their isolated and mixed transport across the human-derived epithelial gastric NCI-N87 and intestinal Caco-2 cells; and evaluate the individual and binary effect of the mycotoxins, in terms of additive, antagonistic and synergistic toxicity towards Caco-2 cells and HepG2 cells.

III.3.1 METHOD DEVELOPMENT AND VALIDATION

Matrix calibration curves were performed to compensate matrix effects on the analytes response. The linearity of the chromatographic method was determined by preparing matrix-matched calibration curves using 6 calibration points ranging from 3.0 to 200 µg/kg in raw/cooked meat and herbs, 1.0 to 100 µg/L in the bioaccessible digested fractions, and 1.0 to 600 µg/L in the transported ones.

113

CHAPTER III. RESULTS AND DISCUSSION

Good linear response was observed for all mycotoxins with coefficients of determination (R2) > 0.9901 in the cooked chicken breast. In the case of herbs and digested matrices, a good linear correlation (R2 > 0.9905 and 0.9914) was obtained for all mycotoxins, except DON and ZEN. It was not possible to validate the method for these two mycotoxins in herbs and DON in the digested matrices. In the case of the transport assay, the four studied mycotoxins had a good linear response with R2 > 0.9916. The LODs and LOQs ranged from 0.5 – 1.25 and 1.5 – 30 µg/kg and 1.5 – 3 and 3.5 – 5 µg/kg for cooked chicken breast and herbs, respectively. Lower LODs and LOQs were achieved in the matrix in vitro digestion and transport experiments ranging from 0.25 – 1 and 0.5 – 5 µg/L (Table III.7 A-C). The percentages of recovery were higher than 70% for all matrices, except FB1 and FB2 in the cooked chicken breast that showed recoveries lower than 60% at the lower spiked levels. The relative standard deviation (%RSD) values for inter-day and intra-day precision were lower than 15% for all matrices (Table III.7 A-C). Therefore, the validated methods were further applied to confirm the absence of mycotoxins in herbs and chicken breast muscle. These methodologies were further applied in cooking, digestion, and transport assays.

114

CHAPTER III. RESULTS AND DISCUSSION

115

CHAPTER III. RESULTS AND DISCUSSION

116

CHAPTER III. RESULTS AND DISCUSSION

117

CHAPTER III. RESULTS AND DISCUSSION

Table III.7 – (continued)

(C) IN VITRO TRANSPORTED FRACTION Spiked level Recovery (%) Intra-day LOD* LOQ* Mycotoxins (µg/L) (Inter-day. %RSD) (%RSD) (µg/L) (µg/L) 0.75 DON 1 89.2 (4.3) 5.32 1 (3.3) (2.53) 5 92 (6.2) 6.5 10 95.7 (4.5) 5.6 AFB1 1 93.9 (7.85) 2.77 0.75 (2.4) 1 (3.2) 5 93 (0) 6.48 10 89 (4.07) 3.73 0.75 FB1 1 98.2 (6.88) 2.78 1 (1.4) (1.04) 5 92.1 (8.46) 6.35 10 99.6 (2.75) 4.4 OTA 1 98.4 (8.49) 4.95 0.5 (1.24) 1 (2.5) 5 99.1 (3.91) 3.46 10 97.2 (3.5) 3.64 *LOD and LOQ values given in brackets are expressed in nM.

III.3.2 IMPACT OF CULINARY PRACTICES

In this study a total of eight cooking methods including microwaving (Mw), microwaving with bag (MwBg), microwaving with a commercial mixture of herbs (MwH), microwaving with bag and the mixture of herbs (MwHBg), oven cooking (Ov), oven cooking with bag (OvBg), oven cooking with the herbs (OvH), and oven cooking with bag and the herbs (OvHBg) were tested to understand the stability of mycotoxins after cooking. The impact of cooking on the 10 studied mycotoxins is presented in Figures III.19A and Figure III.19B. In general, T2 was highly stable to all cooking methods, with percentages after cooking ranging from 90.8 to 97.1% in microwave-based cooking methods, and 85.5 to 98.9% in oven-based cooking methods; whereas AFB1, AFG2, and FBs contents were significantly (p<0.05) decreased by all cooking methods, with remaining percentages after cooking of 50.7-81.6% and 59.1-78.6% (AFB1), 46.2-84.6% and 55.5-82.2% (AFG2), 51.9-60.5% and 56.7-72.2% (FB1), and 58.9-70.1% and 62.4-82.6% (FB2) in microwave-based and oven- based cooking methods, respectively. Also, all microwaving methods have a significantl impact on OTA’s content (p<0.05) with remaining percentages ranging from 60.9 to 74.9% after cooking. Regarding the other mycotoxins a reduction rate was generally verified, although not always statistically significant.

118

CHAPTER III. RESULTS AND DISCUSSION

Figure III.19 – The impact of oven cooking (A) and microwaving (B) on impact mycotoxins. Data is expressed as the percentage (%) of mycotoxin that remained after cooking in the presence or absence of herbs and/or roasting bag. The dashed line relates to the initial concentration quantified in raw samples, normalized to 100%; Data expressed as mean ± standard deviation or as median (minimum-maximum), (n =6). Ov – oven cooking; OvBg – oven cooking with bag; OvH – oven cooking with herbs; OvHBg – oven cooking with bag and herbs; Mw – microwaving; MwBg – microwaving with bag; MwH – microwaving with herbs; MwHBg – microwaving with bag and herbs; Columns with different letters in each mycotoxin differ significantly (p<0.05).

To the best of our knowledge only two reports were found on the impact of cooking of studied mycotoxins - AFB1 and AFB2 (164) and OTA (6) - in meat products, having no information in literature concerning the impact of cooking meat on the stability of the other mycotoxins analysed in this study. Literature shows some contradictory results reporting mycotoxins either as highly stable to high temperatures or also as thermo-labile in other studies (160-162). Herein, microwaving (Mw) and oven cooking (Ov) themselves have low to moderate impact on most mycotoxins content, with FBs being the most cooking-labile, followed by AFs, OTA, DON, ZEN, and T2 as the most heat-stable. Mycotoxins have been described as resistant to thermal-processing treatments (159- 162) although their degradation is dependent on temperature, time and moisture conditions during processing (161). Literature reports higher degradations of AFs, DON, T2, OTA, and FBs in food products with higher moisture content, whereas ZEN stability seems not to be influenced by moisture content (159-163). The samples herein evaluated had high water 119

CHAPTER III. RESULTS AND DISCUSSION

content: 75% for raw samples; 66.8 to 73.1% for microwaved samples (Mw, MwBg, MwH, MwHBg); and 63.7 to 69.6% for oven-cooked samples (Ov, OvBg, OvH, OvHBg), so a higher degradation of mycotoxins would be expected. Trichothecenes were poorly reduced by Mw (350 w, 0.45s) and Ov (190 ºC, 5min) with reductions ranging from 18 to 27% for DON and from 8.6 to 12.4% for T2, although higher degradation was expected. Time, temperature and pH conditions determine the extent of DON degradation, with alkaline condition (pH 10) and high temperatures (100-200 ºC) completely degrading this mycotoxin (369). The stability of DON under various temperatures (150 to 200 ºC) and times (5 to 20 min) in different matrix models (sugar, starch and protein) was evaluated by Bretz and colleagues (370), reporting that the protein model (N-R-acetyl-L-lysine methyl ester) contributed with the fastest degradation. Pyrolysis and polymerization reactions were pointed as possible reactions for DON degradation and formation of degradation products (370); in the case of T2, higher degradations were reported during extrusion-cooking (369). On the other hand, the moisture and pressure/shear force parameters had a major impact on T2 degradation, with temperature not being so determinant (163). In the same way, the aflatoxins were also considerably stable after Mw and OV with maximum losses of 21.4%. Furtado and colleagues (164) verified reductions of AFB1 and AFB2 ranging from 15 to 30 % after frying pork bellies (171 ºC, 3 min) and broiling pork chops in an electric oven to an internal temperature of 76 ºC. However, a complete degradation of AFB1 was verified after treating matrix soya protein at 150 ºC for 15 min (371). The availability of water in the matrix and the limited stability of a hydrolysed lactone of AFB1 were pointed as the main factors affecting its degradation. Also, AFB1 degradation may also happen through the opening of the lactone ring at higher temperatures (> 100 ºC) and at alkaline conditions, as well as by decarboxylation of o-coumaric acid in more drastic thermal conditions (371). FBs were the most affected group of mycotoxins, with Mw and Ov reducing their contents up to 42%. Similar reductions (50%) were observed after frying (190 ºC) corn chips; extrusion-cooking (160 ºC) corn grits; and roasting (218 ºC) corn meal (160). During food processing, fumonisins can bind different components within the food matrix or react with them contributing for the formation of the so called “hidden fumonisins” that are not detected by conventional methods (162). Seefelder and colleagues (372) demonstrated that FB1 was able to bind to protein after thermal treatment. In this study only the mycotoxins present in their free form, i.e. un-bounded, were detected, thus it cannot be excluded the possibility that FBs reductions found here may result from molecular degradation or bound to proteins/peptides or other compounds.

120

CHAPTER III. RESULTS AND DISCUSSION

OTA was quite stable to Mw and Ov, which agrees with Pleadin and colleagues (6) who evaluated the impact of frying (170 ºC, 30 min) and boiling (100 ºC, 30 min) on OTAs stability in meat sausages concluding that this toxin was poorly lost after cooking with reduction rates of 12.6% and 7.4%, respectively, and only increasing time of cooking to 60 min resulted in significant reductions (75.8%). The decrease of OTA content after cooking may be due to the amine breakdown at high temperatures, or due to OTA conversion into L- phenylalanine and ochratoxin-α (371). ZEN was also highly stable to Mw and Ov, with percentages of loss ranging from 8.3 to 17.7%. Baking and roasting may decrease ZEN content, but due to ZEN high thermal- stability (373), and even more considering the short period of time used in the present study (0.45 s and 5 min), ZEN content was poorly affected during cooking. Else, this mycotoxin was highly stable in an aqueous buffered model system to a wide range of temperature (100 to 200 ºC) in the first 30 min, regardless of pH (374). Concerning the additional cooking conditions used in this study, cooking inside the bags (MwBg and OvBg) did not confer any advantage in reducing mycotoxin when compared with Mw and Ov cooking, whereas adding herbs significantly affected (p<0.05) mycotoxins stability and reduced their contents when compared with the other cooking methods. Only for FB2, the use of herbs was not effective in reducing its content. Combining the use of bags with the addition of herbs significantly affected some mycotoxins either during MwHBg and OvHBg, but it is interesting to notice that, despite the different cooking-stability verified among mycotoxins, MwHBg and MwH cooking treatments did not significantly differ for all mycotoxins. Although not as evident as in microwaving, the use of herbs, with (OvHBg) or without (OvH) the use of bags, seems also to be the strongest factor to reduce mycotoxins’ content. This shows that the addition of herbs has a stronger effect on mycotoxins stability than the use of bags, i.e. the use of bag while microwaving with herbs did not significantly differ from microwaving only with herbs. This may suggest that the addition of herbs prior to oven cooking or microwaving chicken breast helps reduce mycotoxins content. The reduction of mycotoxins content after cooking with herbs (MwH, OvH, MwHBg, and OvHBg) likely results from the mitigating effect of herbs toward mycotoxins. Herbs are rich sources of powerful antioxidants (375), commonly used in home cooking and have been revealed to chemically interact with mycotoxins (159, 161). The removal of mycotoxins by botanicals is presently being explored and is usually preferred over chemical treatments (161) and has been considered as a safe methodology to remove aflatoxins from food without causing harm to humans and animals (376). Some researchers reported the potential of ajwain spice (carom) Trachyspermum ammi (L.), commonly used in Asian cuisine, to degrade aﬂatoxins (306, 377). Additionally, extracts of medicinal plants, Ocimum

121

CHAPTER III. RESULTS AND DISCUSSION

tenuiflorum mitigate aflatoxins, being suggested as a safe additive, which not only inhibits the aﬂatoxin production throughout fungal growth inhibition but also degrades the existing toxin (378), using aqueous extracts of Justicia adhatoda (adulsa, also known as Adhatoda vasica Nees) and leaf extracts of Corymbia citriodore (eucalyptus) also degraded AFB1 (> 90 %) (379). The probable mechanism of degradation of aflatoxin relies on modification of lactone ring in aflatoxin structure (306). No literature information was found concerning the effect of adding plant extracts on the other studied mycotoxins, thus further investigation is required to properly evaluate the impact of adding herbs and other phytochemicals on mycotoxins stability.

III.3.2.1 EXPLORATORY AND STATISTICAL ANALYSIS

For a better interpretation of the full data set concerning the impact of cooking on the stability of mycotoxins, an agglomerative hierarchical clustering (AHC) analysis represented as heatmap was performed, allowing a rapid and effective visualization of the similarities and differences between cooking methods and mycotoxins stability. The heatmap representation of the impact of cooking on mycotoxins stability is shown in Figure III.20. The cooking methods were grouped into three clusters (horizontal dendrogram): the first formed between cooking methods without the addition of herbs (Mw, MwBg and OvBg), followed by clustering of Ov and OvH, and finally, the most dissimilar cluster with the other cooking methods that used herbs (MwH, MwHBg and OvHBg). This graphical form of representation emphasises that the addition of herbs impacts the mycotoxins stability during cooking, although OvH not being clustered with MwH, MwHBg, and OvHBg. OvH was the least effective method with the addition of herbs on reducing mycotoxins when comparing with Ov and microwave-based cooking methods, which justifies its clustering with Ov. Moreover, the lower mycotoxin content after cooking (greener colouration in the image) was observed for MwH, MwHBg, and OvHBg for almost all mycotoxins when comparing with other cooking methods indicating that the addition of herbs before cooking favours mycotoxins reduction. Furthermore, the clustering between Mw and MwBg, and the similar Euclidean distances between Ov and OvBg suggest that the use of bag while cooking may not be advantageous on reducing mycotoxins content. On the whole, Ov appears to be the least effective (more reddish) cooking method on mycotoxins stability, whereas MwH seems to be the most effective one (greener).

122

CHAPTER III. RESULTS AND DISCUSSION

Figure III.20 – A - Cluster heatmap representation of all eight cooking methods concerning mycotoxins’ percentage after cooking; Data expressed as mean or as median for samples following normal or non-normal distribution, respectively. Autoscaling was chosen for data scaling and Euclidean distance and Ward linkage were selected to establish the clusters; Mw – microwaving; MwBg – microwaving with bag; MwH – microwaving with herbs; MwHBg – microwaving with bag and herbs; Ov – oven cooking; OvBg – oven cooking with bag; OvH – oven cooking with herbs; OvHBg – oven cooking with bag and herbs.

In the case of mycotoxins’ grouping by similarities (vertical dendrogram) their cooking stability seems to be structurally-related, as FBs, thricothecenes (DON and T2), and AFs (except AFB2) appear in three different clusters. The first cluster was formed between mycotoxins FB1, FB2 and OTA with all three being more stable to Ov and OvH and less stable (showing a greener colouration) to MwH and MwHBg. The second cluster was formed between three of four AFs (AFB1, AFG1, and AFG2) and ZEN. Finally, the most dissimilar mycotoxins (i.e. those with higher Euclidean distances between them) were T2, DON and AFB2 forming the third cluster, with DON and AFB2 being more similar, with higher stability to Mw, MwBg, and OvBg, and T2 being more stable to cooking methods with added herbs (except OvH). To study the relationship between percentage of mycotoxins remaining after cooking and the type of cooking applied (i.e. Mw, Ov, use of bag and/or herbs), a PLS regression was performed based on mycotoxins stability prediction (collectively (for all mycotoxins) and individually) from the type of cooking used (Table III.8).

123

CHAPTER III. RESULTS AND DISCUSSION

Table III.8 – Multivariate Partial Least Square (PLS) regression between percentage of mycotoxins after cooking (X – variables) and type of cooking applied (i.e. microwaving or oven cooking, presence or absence of bag, and/or herbs addition) (Y – variable). Q² R²X R²Y RMSE

All mycotoxins 0.254 0.333 0.308 0.778

Isolated DON -0.327 0.333 0.417 0.714 T2 0.230 0.333 0.502 0.706 AFB1 0.577 1.000 0.848 0.311 AFB2 0.444 0.667 0.789 0.269 AFG1 0.626 1.000 0.878 0.334 AFG2 0.524 1.000 0.882 0.163 FB1 -1.237 1.000 0.438 0.628 FB2 -1.284 1.000 0.412 0.648 OTA -0.271 1.000 0.627 0.571 ZEN -1.672 1.000 0.467 0.561 Q² - Cumulative predictive variation from internal cross-validation; R²X – cumulative explained variation of X explained in terms of sum of squares; R²Y – cumulative explained variation of Y explained in terms of sum of squares; RMSE – root mean square error.

A poor regression model (Q2 < 0.500, R2X and R2Y < 0.700, with high RMSE values) was obtained evaluating all mycotoxins collectively. However, when evaluating individually, an effective multivariate PLS regression model, with good predictive ability (Q2> 0.500, R2X and R2Y > 0.700, with low RMSE values), was able to confirm the influence of herbs addition on the percentage of 3 mycotoxins (AFB1, AFG1, and AFG2) after cooking, regardless of type of cooking method used. In the case of AFB2 the predictive ability of the regression model was not as effective as to the other aflatoxins, with a Q2 value below 0.500 (Q2=0.444). The importance of Y-variables (types of cooking) for the model projection (expressed as VIP values) and their standardized coefficients were determined (Figures III.21 1) to 4)): The use of herbs was classified as a VIP (with values >1) for AFB1, AFG1, and AFG2 and as a moderately VIP for AFB2 (VIP=0.976). These results show that adding herbs offers a negative correlation with % of mycotoxin after cooking, reducing its content.

124

CHAPTER III. RESULTS AND DISCUSSION

Figure III.21 – Cooking impact, translated as estimated standard regression coefficient, for the selected AFB1 (1), AFB2 (2), AFG1 (3), and AFG2 (4), modelled by PLS with 95 % confidence interval (Variable importance for the projection (VIP) >1 are represented in red bold and moderately 0.8< VIP <1 are in black bold); Bag-N – without bag; Bag-Y – with bag; Herbs-N – without herbs; Herbs-Y – with herbs; MW – microwaving.

III.3.3 MYCOTOXINS’ BEHAVIOUR THROUGH GASTROINTESTINAL TRACT

Gastric and duodenal bioaccessibility of T2, AFB1, AFB2, AFG1, AFG2, FB1, FB2, OTA and ZEN from microwaved and oven-cooked chicken breast muscle are summarized in Tables III.9 and Table III.10, respectively. The bioaccessibility of DON was not possible to evaluate as described in Section III.3.1. Notwithstanding, an overall increasing tendency (p<0.05) was observed over digestion increasing the bioaccessibility of mycotoxins from gastric to duodenal digestion phase. The global percentages of bioaccessibility ranged from

125

CHAPTER III. RESULTS AND DISCUSSION

The type of cooking used (i.e. Mw or Ov) or use of roasting bags seems not to be a determinant factor concerning the mycotoxins’ bioaccessibility, with the type of mycotoxin playing a more important role. Our results differ from those described in literature since higher duodenal bioaccessibility of mycotoxins were reported on cereal-based products (185-187, 189). Herein, the lower bioaccessibility observed may probably result from mycotoxin’ linkage to proteins causing lower availability after digestion. This agrees with the effects observed by adding protein ingredients (milk whey, β-lactoglobulin, and calcium caseinate) to wheat crisp bread that reduced 74-88% mycotoxin bioaccessibility (190) and to the strong tendency of AFB1 and OTA to form covalent bonds with proteins reported by Raters and colleagues (371). In fact, AFB1 and OTA, along with ZEN, were the mycotoxins with the lowest bioaccessibility. A significant increase of mycotoxins bioaccessibility, ranging from 32 to 56%, was observed when comparing MwH and OvH with other cooking methods that did not use herbs (Mw, MwBg, Ov, and OvBg), therefore, the addition of herbs significantly increased the bioaccessibility of mycotoxins, mostly the AFG2 that exhibited the highest percentages of bioaccessibility. Concerning the combined effect on bioaccessibility, MwHBg significantly differed (p<0.05) on AFG2 bioaccessibility in comparison with Mw, exhibiting a 33% increase. In the case of OvHBg, only the bioaccessibility of fumonisins were significantly affected (p<0.05) in comparison with Ov, OvBg, and OvH, increasing the percentage of FB1 and decreasing the percentage of FB2.

126

CHAPTER III. RESULTS AND DISCUSSION

127

CHAPTER III. RESULTS AND DISCUSSION

128

CHAPTER III. RESULTS AND DISCUSSION

Despite this later result, the percentages of bioaccessibility of all other mycotoxins did not differ among cooking methods using herbs, with (OvHBg, MwHBg) or without (OvH, MwH) the use of bag, suggesting that the herbs addition is the most important factor affecting mycotoxins bioaccessibility. However, dietary fibers, as galactomannans, glucomannans, glucans, and cellulose were found to decrease the bioaccessibility of mycotoxins (380). So, concerning the common use of herbs as seasoning ingredients and noticing that in the case of mycotoxins these seem to have an impact in both cooking- stability and bioaccessibility, these type of natural seasonings must be subject of further evaluation concerning their impact during cooking and consequent bioaccessibility of several compounds (undesirable or desirable).

III.3.3.1 EXPLORATORY AND STATISTICAL ANALYSIS

An AHC analysis was also performed to interpret the impact of cooking on mycotoxins’ bioaccessibility (Figure III.22).

Figure III.22 – A - Cluster heatmap representation of all eight cooking methods concerning mycotoxins’ bioaccessibility. Data expressed as mean or as median for samples following normal or non-normal distribution, respectively. Autoscaling was chosen for data scaling and Euclidean distance and Ward linkage were selected to establish the clusters; Mw – microwaving; MwBg – microwaving with bag; MwH – microwaving with herbs; MwHBg – microwaving with bag and herbs; Ov – oven cooking; OvBg – oven cooking with bag; OvH – oven cooking with herbs; OvHBg – oven cooking with bag and herbs;

The cooking methods are grouped into three clusters (horizontal dendrogram) with the most dissimilar cluster composed by MwH and MwHBg. Microwaving with herbs show an 129

CHAPTER III. RESULTS AND DISCUSSION

orange to red colouration for all mycotoxins meaning bioaccessibility above the average, which is interesting considering that in the previous heatmap (Figure III.20) these two cooking methods were the most effective in decreasing mycotoxins content. Even though mycotoxins are in lower amounts after cooking, these may be highly bioaccessible after digestion. Similarly, oven-based methods that reduced mycotoxins content during cooking, show a wide variance on bioaccessibility. Thus, concerning the influence of type of cooking on mycotoxins bioaccessibility, MwH and MwHBg are the cooking methods that mostly contribute for higher mycotoxins bioaccessibility, while Mw is the cooking method allowing minor bioaccessibility. FBs along with OTA were the most cooking-labile mycotoxins described before, however, even though present in lower amounts they show an intermediate bioaccessibility. Furthermore, the lower AFB1 and ZEN bioaccessibility (green colouration) in cooking methods without herbs are confirmed as previously discussed (Tables III.9 and Table III.10, and Figure III.22). Mycotoxins grouping by similarities concerning their bioaccessibility (vertical dendrogram), though not as evident as in Figure III.20, seems also to be structurally- related, which agrees with previous results suggesting that the type of cooking is not a determinant factor for mycotoxins’ bioaccessibility, but the type of mycotoxin plays a more prominent role. Interestingly, while Figure III.20 suggests a high decrease of AFB1, AFG1, AFG2 during cooking with herbs (MwH, MwHBg, and OvHBg), in Figure III.22 the same mycotoxins show higher bioaccessibility (red colouration) in those methods. Moreover, AFG2 was described as the most bioaccessible mycotoxin in all cooking methods with average bioaccessibility ranging from 56.1 to 92.5%, and > 80% in cooking methods with herbs. It is interesting to notice that this mycotoxin was highly affected by cooking methods with herbs (percentages after cooking < 53%), however this lower content seems to be highly available after in vitro digestion. To study the relationship between percentage of mycotoxins bioaccessibility and the type of cooking applied (i.e. Mw, Ov, use of bag and/or herbs), a PLS regression was performed based on mycotoxins bioaccessibility (collectively (for all mycotoxins) and individually) from the type of cooking used (Table III.11).

130

CHAPTER III. RESULTS AND DISCUSSION

Table III.11 – Multivariate Partial Least Square (PLS) regression between percentage of mycotoxins percentage of bioaccessibility (X – variables) and type of cooking applied (i.e. microwaving or oven cooking, presence or absence of bag, and/or herbs addition) (Y – variable). Q² R²X R²Y RMSE

All mycotoxins 0.167 0.333 0.234 0.819

Isolated DON T2 0.230 0.333 0.574 0.706 AFB1 0.577 1.000 0.807 0.411 AFB2 0.444 0.667 0.660 0.545 AFG1 0.626 1.000 0.882 0.321 AFG2 0.524 1.000 0.794 0.424 FB1 -1.237 1.000 0.327 0.767 FB2 -1.284 1.000 0.305 0.78 OTA -0.271 1.000 0.626 0.572 ZEN -1.672 1.000 0.293 0.787 Q² - Cumulative predictive variation from internal cross-validation; R²X – cumulative explained variation of X explained in terms of sum of squares; R²Y – cumulative explained variation of Y explained in terms of sum of squares; RMSE – root mean square error.

A poor regression model (Q2 < 0.500, R2X and R2Y < 0.700) was obtained evaluating all mycotoxins collectively. However, when evaluating individually, an effective multivariate PLS regression model, with good predictive ability (Q2> 0.500, R2X and R2Y > 0.700, with low RMSE values), indicates an influence of herbs addition on the percentage of bioaccessibility of AFB1, AFG1, and AFG2, regardless of cooking method used (Figure III.23 1) to 4)), validating that indeed adding herbs offers a positive correlation with % of mycotoxin bioaccessibility, increasing aflatoxins bioaccessibility. In the case of AFB2 the predictive ability of the regression model was not as effective as to the other aflatoxins, with a Q2 value below 0.500 (Q2=0.444).

131

CHAPTER III. RESULTS AND DISCUSSION

Figure III.23 – Bioaccessibility, translated as estimated standard regression coefficient, for the selected AFB1 (1), AFB2 (2), AFG1 (3), and AFG2 (4), modelled by PLS with 95 % confidence interval (Variable importance for the projection (VIP) >1 are represented in red bold and moderately 0.8

The importance of Y-variables (types of cooking) for the model projection (expressed as VIP values) and their standardized coefficients were determined (Figures III.23 1) to 4)): The use of herbs was classified as a VIP (with values >1) for AFB1, AFG1, and AFG2. These results show that adding herbs offers a positive correlation with % of mycotoxin bioaccessibility, promoting their content after in vitro digestion.

132

CHAPTER III. RESULTS AND DISCUSSION

III.3.4 MYCOTOXINS GASTRIC AND INTESTINAL ABSORPTION

In this study four mycotoxins – AFB1, DON, OTA, and FB1 – from four different chemical groups were selected to evaluate their transport rate isolated or in mixture across gastric NCI-N87 and intestinal Caco-2 monolayers mimicking the gastric and intestinal epithelium where mycotoxins absorption may occur.

III.3.4.1 QUALITY CONTROL OF MONOLAYERS

To confirm that the transport study was performed under recommended conditions two different “quality control” measurements were considered to check the monolayer integrity: i) measurement of TEER; and ii) determination of mycotoxins cytotoxicity in proliferating cells at the concentrations studied. TEER values of NCI-N87 and Caco-2 monolayers were in agreement with values previously reported by Lemieux and colleagues (229) (> 500 Ω cm2 ) and Melo and colleagues (381) (> 1000 Ω cm2) for NCI-N87 and Caco-2, respectively. No significant differences (p<0.05) were observed on TEER values for both cell lines, measured before (0h) and after transport experiments (3h), after exposure to mycotoxins, individually or in mixture (MIX) (Figure III.24 A-D). A decrease in TEER values could mean an increase in permeability of tight junctions (382) due to cells detachment. Moreover, DON (1.69 µM), AFB1 (1.60 µM), FB1 (0.69 µM), OTA (1.24 µM), isolated or in mixture, did not exert significant cytotoxic effect in proliferating NCI-N87 and Caco-2 cells after 3h exposure, not influencing cell viability (Figure III.24 E-F) measured by the MTT assay and expressed as % of control. These results indicate that the monolayer integrity of differentiated NCI- N87 and Caco-2 cells was not compromised during transport assays.

133

CHAPTER III. RESULTS AND DISCUSSION

Figure III.24 – Transepithelial electrical resistance (TEER) values (Ω cm2) of DON, AFB1, FB1, OTA, and their mixture (MIX) before (0h) and after (3h) the transport experiment for: NCI-N87 cells in the apical  basolateral direction (A) and in the basolateral  apical direction (B); and Caco-2 cells in the apical  basolateral direction (C) and in the basolateral  apical direction (D). Figures III.24E and III.24F show the % of viability of proliferating NCI-N87 and Caco-2 cells, respectively, after 3h exposure to mycotoxins isolated and in mixture. Data expressed as mean ± SD of 3 independent experiments (n=3). AFB1 – Aflatoxin B1; DON – deoxynivalenol; FB1 – Fumonisin B1; MIX – mixture of four mycotoxins; and OTA – Ochratoxin A. Columns with different letters in each ACD concentration differ significantly (p<0.05).

III.3.4.2 TRANSPORT STUDIES

The bidirectional transport of isolated DON, AFB1, FB1, OTA, as well as their mixture was assessed in NCI-N87 and Caco-2 monolayers through a concentration gradient. These models mimic the bidirectional transport occurring in gastric and intestinal absorption and allow the calculation of the uptake and efflux ratios. Transport processes across absorptive epithelia may occur through several routes: passive (transcellular and/or paracellular), active transcellular (transporter-mediated) or transcytosis (329). Previous research on the mycotoxins mechanism and routes of intestinal uptake propose AFB1 and DON as passively transported (DON using a paracellular route), OTA also passively transported through simple diffusion, and the FB1 entero-hepatic circulation was 134

CHAPTER III. RESULTS AND DISCUSSION explained by its interaction with cholesterol or bile salts facilitating its intestinal transport (239, 243, 383, 384). The present study focused on the passive transport of the four mycotoxins across gastric and intestinal monolayers. The results concerning transport of each mycotoxin isolated and in mixture, expressed as mass transported over time, are shown in Figures III.25 (gastric transport) and Figure III.26 (intestinal transport) across NCI-N87 and Caco-2 monolayers, respectively. Different transport rates were observed concerning the type of mycotoxins, the type of transport (isolated or mixed), the cell line, and the flux direction.

Figure III.25 – The percentage of DON, AFB1, OTA, and FB1 transferred to the receiver compartment over 180 min across monolayers of NCI-N87 in the apicalbasolateral and basolateralapical directions when transported isolated (green dots and lines) or in mixture (orange dots and lines). The dots with SD represent the experimental values obtained in this study, while the line links the theoretical values obtained from the equation showed in Chapter II.4.2. Data expressed as mean ± SD of 3 independent experiments (n=3). AFB1 – Aflatoxin B1; DON – deoxynivalenol; FB1 – Fumonisin B1; MIX – Mixture; and OTA – Ochratoxin A. 135

CHAPTER III. RESULTS AND DISCUSSION

Regarding the gastric absorption (Figure III.25), all mycotoxins were transported, either isolated or in mixture through NCI-N87 cells in the apical to basolateral (AB) direction, except the FB1 that was not transported in mixture; whereas, only DON and AFB1 were transported in the opposite direction, either isolated or in mixture. In the case of intestinal absorption, FB1 was not transported in any direction, OTA was absorbed in AB direction and not transported in the opposite direction, and DON and AFB1 were transferred in both directions through Caco-2 monolayers; moreover, for some mycotoxins different transport profiles were observed between isolated and mixed transport (Figure III.26).

Figure III.26 – The percentage of DON, AFB1, and OTA transferred to the receiver compartment over 180 min across monolayers of Caco-2 in the apicalbasolateral and basolateralapical directions when transported isolated (green dots and lines) or in mixture (orange dots and lines). The dots with SD represent the experimental values obtained in this study, while the line links the theoretical values obtained from the equation shown in Chapter II.4.2. Data expressed as mean ± SD of 3 independent experiments (n=3). AFB1 – Aflatoxin B1; DON – deoxynivalenol; MIX – Mixture; and OTA – Ochratoxin A.

136

CHAPTER III. RESULTS AND DISCUSSION

Individually, the transfer of DON isolated or in mixture to the receiver compartments was linear with time in both cells and transport directions (AB and basolateral to apical (BA)). In NCI-N87 cells, DON was similarly transported isolated or in mixture in both directions, achieving maximum percentages around 8% (Figure III.25). On the other hand, at intestinal level, a faster bidirectional transport of DON was observed in mixture (orange lines), mostly in the AB direction achieving maximum percentages of 14%, while in the same direction the isolated transport showed maximum percentages around 8%. This suggests that the presence of other mycotoxins may promote intestinal transport of DON (Figure III.26). AFB1 transport rate was found non-linear being rapidly transferred to the receiver compartment and decreasing its rate with time until stabilization, either when transported isolated or in mixture. This phenomenon occurred in both epithelium and both transport directions, increasing rapidly in the first 60 min and more slowly until 180 min. AFB1 basolateral uptake isolated or in mixture differed across NCI-N87 monolayers, showing higher percentage of transport over time in the mixed transport; while, in the opposite direction, the isolated transport was faster, achieving higher transport rates. Concerning Caco-2 transport, AFB1 was equally transported either isolated or in mixture in the AB direction, however, in the opposite direction a faster and higher transport was observed in mixture, suggesting that the intestinal transport of AFB1 in the presence of other mycotoxins is beneficiated in the efflux direction. OTA transport across NCI-N87 cells was similar if transported isolated or in mixture, and only happened in the AB direction; however, OTA transport isolated or in mixture highly differed across Caco-2 monolayers, being poorly transported isolated and more rapidly transferred when in mixture. FB1 was only poorly transported at gastric level regarding basolateral uptake with a maximum flux of 1%, not being transported in mixture, neither in the opposite direction nor across Caco-2 monolayers in both directions. Beyond expressing the amount of mycotoxin transported over time, it was also determined the apparent permeability (Papp), expressed as cm/s, to accurately predict the absorption of mycotoxins via gastrointestinal tract. As AFB1 and OTA were rapidly absorbed at the beginning with the transport rate decreasing with time, it can be inferred that the sink conditions were not verified during the experiment contributing for a shallower concentration gradient between donor and receiver compartments, instead of a linear fitting between mycotoxin content in the donor and receiver compartment. Therefore, following the advice of the method authors, the permeability values were calculated using the equation mentioned in Chapter II.4.2 for nonlinear curve fitting developed for non-sink conditions

(231, 329). Tavelin and colleagues (329) reported that Papp obtained with non-sink

137

CHAPTER III. RESULTS AND DISCUSSION

conditions are in better accordance with the real permeability coefficients of the epithelial cell monolayer in vivo. Thus, Figure III.27 shows the Papp values calculated from the curves of the transport of isolated mycotoxins present in the Figures III.25 and Figure III.26.

Figure III.27 – Apparent permeabilities (Papp) of isolated DON, AFB1, FB1, and OTA transported across NCI- N87 (gastric) and Caco-2 (intestinal) cells over 3h in both AB and BA directions. Grey and white arrows indicate AB or BA transport, respectively, and their thickness mean higher or lower permeability; the red “X” indicates no transport across the monolayer. Data expressed as mean ± SD of 3 independent experiments (n=3). AFB1 – Aflatoxin B1; DON – deoxynivalenol; FB1 – Fumonisin B1; and OTA – Ochratoxin A. Values with different letters or numbers differ significantly (p<0.05) from the given mean (lowercase letters for AB gastric transport, uppercase letters for BA gastric transport, roman numerals for AB intestinal transport, and arabic numbers for BA intestinal transport). *Significant difference (p<0.05) between AB and BA permeabilities for the same mycotoxins in the cell line. # Significant difference (p<0.05) between gastric and intestinal permeabilities for the same compound.

As expected by the isolated transport rates observed in Figures III.25 and Figure III.26, in general, AFB1 exhibits the highest Papp, followed by DON, and then with the lowest permeabilities are FB1 and OTA in those directions to which transport was observed (apart from AB flux of OTA in gastric monolayers) (Figure III.27). Comparing the different mycotoxins, the gastric Papp in the AB direction was significantly (p<0.05) higher for AFB1 and OTA in comparison with DON and FB1, while all intestinal transport in AB direction differed among mycotoxins with AFB1 having the highest and OTA the lowest Papp. From those mycotoxins that were transported in the opposite direction (BA) the permeabilities obtained also significantly differed (p<0.05) in both cell lines with AFB1 having higher Papp rather than DON which was verified in both cell lines. Concerning the transport of the four mycotoxins in mixture, results showed significant differences (p<0.05) in the Papp as can be seen in Table III.12.

138

CHAPTER III. RESULTS AND DISCUSSION

139

CHAPTER III. RESULTS AND DISCUSSION

At the gastric level, the co-transport of mycotoxins significantly decreased the Papp of

AFB1 and blocked the AB flux of FB1; while for the intestinal flux an increase of Papp was observed for DON and OTA when simultaneously transported (Table III.12). The mass balance values in both transport directions (calculated as described in Chapter II.3.5) ranged from 70% to 116% for NCI-N87 and Caco-2 cells, supporting acceptable approximation of the Papp values (231). Artursson and colleagues (385) proposed that Papp

-6 values greater than 1x10 cm/s indicate high permeability coefficients, thus the Papp values obtained for mycotoxins suggest that AFB1, DON, and OTA were efficiently absorbed in NCI-N87 and Caco-2 cells either isolated or in mixture, excepting OTA isolated in intestinal uptake. From our results it seems that the passive bidirectional transport evaluated herein might be pH-dependent with transport rates differing between gastric and intestinal transport, if we consider the pH-partition hypothesis in which ionisable compounds diffuse through biological membranes primarily in their non-ionized form (386). In gastric transport the pH difference between apical (pH 3.0) and basolateral (pH 7.4) compartments seem also to influence the transport direction of FB1 and OTA being only transported in the AB direction.

As these mycotoxins are weak acids (pKa FB1/OTA 3.49/4.40), and according to

Henderson-Hassel Balch equation, a pH near its pKa suggests that 50% of the molecule is non-ionized, its capacity to pass through the cell membranes is thus higher on the acidic than on the neutral side of the monolayer (387, 388). Moreover, Afsah-Hejri and colleagues (389) reviewed that in animals OTA is usually absorbed in the stomach due to its lipid- soluble, nonionized, and acidic properties, which is in agreement with the present results as OTAs transport was higher across NCI-N87 than Caco-2 monolayers. Similarly, as intestinal transport through Caco-2 monolayers is performed at physiological pH (7.4) in both sides, it was not expected FB1 to be transported in any direction by passive non-ionic transcellular diffusion because at duodenum pH the FB1 is mainly negatively charged (387). There are no reports on gastric absorption of FB1 to discuss our results, but the absence of transport of FB1 across Caco-2 monolayer is in agreement with De Angelis and colleagues (241). Also, EFSA reports describe a poor absorption of FB1 from animal experiments, which agrees with the poor FB1 found in the present study (lower than 2%, and happening only in gastric transport) (220). In the case of OTA, concerning the pKa of phenolic hydroxyl group of 7.1 it would be expected some transport in the BA direction at gastric level, as well as bidirectional transport of this mycotoxin across Caco-2 monolayers, but it was only observed a weak AB transport of OTA. Bidirectional transport of OTA using pH 7.4 was reported by Berger and colleagues (239), with low transport levels in AB direction and high transport in the opposite side. However, authors verified that the addition

140

CHAPTER III. RESULTS AND DISCUSSION of BSA in the basolateral compartment, as performed here, can have the effect of drastically decreasing the transport in BA direction. This is in agreement with EFSA reporting that when OTA reaches the systemic circulation it extensively binds to plasma proteins, e.g. serum albumin, suggesting that only a small fraction remains in its free form (218). The strategy of adding BSA to the basolateral compartment, despite its influence on the transport results, increases the reliability of the model as the serum contains equally high levels of proteins. In the same study, Berger and colleagues also demonstrated that decreasing the pH in the apical side to pH 6.0 increased the AB flux. In addition, Qi and colleagues (390) reported that, considering both pKa’s (4.4 and 7.1) of OTA, most OTA in nature is expected to be present as organic anion, supporting the thesis of transmembrane transporters as significant pathways for the uptake and efflux of OTA. Several transporters have been described to mediate OTAs uptake and efflux across epithelial cells as the organic anion- transporting polypeptides 1A2, 1B1, 2B1 (OATP1A2, OATPA1B1, OATP2B1), which mediate the uptake of OTAs and the breast cancer resistance protein (BCRP) and the multidrug resistance protein 2 (MRP2) transporters known to mediate the efflux of OTA out of the epithelial cells back into lumen (198, 241, 390, 391). The limited intestinal absorption of OTA was justified by the presence of MRP2 transport at the apical side of Caco-2 cells (392). Thus, the use of physiological pH, the addition of BSA in the basolateral, and the presence of efflux transporters may justify the low transport in AB direction and the absence of transport in the BA direction in Caco-2 cells. Concerning AFB1 and DON bidirectional transports, these were found to be insensitive to pH gradients as each mycotoxin was equally transported in both directions (AB or BA) at gastric and intestinal level. There are no literature reports on AFB1 and DON gastric absorption, while for intestinal absorption higher transport rates were reported for AFB1 rather than DON (242, 393). Gratz and colleagues (393) and Mata and colleagues (394)

-6 studied AFB1 absorption across Caco-2 in AB direction reporting lower Papp (~20x10 cm/s) values than those observed herein, probably because our experiment was made under non- sink conditions and the respective equation used to calculate the permeabilities usually

-6 results in higher Papp values (329). Moreover, higher permeability values (105.1±7.89 x10 cm/s), using the non-sink equation, for AFM1 across Caco-2/TC7 cells were reported by

-6 Caloni and colleagues (240). In the case of DON, similar Caco-2 Papp 3.3±0.61 (x10 ) and 5.02x10-6 cm/s were reported by Kadota and colleagues (242) and Sergent and colleagues

(243), respectively. In the BA direction, the Papp value of AFB1 obtained in the present study is similar to the Papp reported by Mata and colleagues (394), as well as the Papp of DON is also in accordance with Sergent and colleagues (243) and Kadota and colleagues (242). These results suggest that AFB1 is more rapidly absorbed than DON, either at gastric or

141

CHAPTER III. RESULTS AND DISCUSSION

intestinal level. The faster intestinal uptake of AFB1 could also be explained by its mediated transport through OAT or OCT transporters for the uptake (395, 396) or MRP transport for the efflux transport of AFB1 (392, 397). AFB1 was found in plasma of lactating dairy cows 5 min after ingestion of contaminated meals suggesting a rapid absorption of AFB1 through gastrointestinal tract of cows (398). Additionally, Jubert and colleagues (399) studied the pharmacokinetics of AFB1 in human volunteers showing a fast absorption of AFB1 into systemic circulation (peak concentration after 1 h). These previous results are in line with the higher transport rate of AFB1 observed in comparison with DON. Sergent and colleagues (243) studied the mechanisms of DON transport across Caco-2 cells suggesting that this mycotoxin can be either transcellular or paracellular passive transported, and that after animal ingestion DON may be absorbed in the proximal part of the small intestine considering its early occurrence in plasma. Moreover, after 30 min of DON oral administration to mice, this toxin has been detected in several organs as spleen, kidney, liver, large intestinal and plasma, and after 1h in the small intestine, suggesting a fast absorption and distribution (400). As discussed until now, the uptake/excretion of toxins is usually evaluated considering a single exposure of these contaminants. However, multiple toxins may be ingested at once and these specific mycotoxins often co-occur in food (48, 401) and increase the toxicity in target organs when in combination (Chapter III.3.5). Thus, transport monitoring of mycotoxins in mixture is of huge importance as, for example, the presence of DON has proved to influence the absorption of several nutrients by inhibiting the uptake of sugars, amino acids, lipids and vitamins (213), and the present study shows that mycotoxin co- transportation may differ from their individual transport (Figures III.25 and Figure III.26, and Table III.12). Other parameters that are usually calculated to evaluate the transport of substances across monolayers are the ratio of uptake (UR) and efflux (ER). An asymmetric transport through cell monolayers (i.e UR or ER higher than 2) usually suggests the involvement of transporters mediating the passage of molecules. As only DON and AFB1 were transported in both flux directions the calculation of uptake/efflux ratios was possible only for these two mycotoxins (Table III.12). A high uptake ratio (1.74) and low efflux ratio (0.57) was observed for DON individually transported through NCI-N87 cells, meaning that this mycotoxin is preferably transported in the AB direction in the stomach, while similar values of UR (1.08) and ER (0.93) were observed in the intestinal model suggesting no preferable AB or BA flux. Individually transported AFB1 showed higher uptake ratios (1.28 and 3.76) than efflux ratios (0.78 and 0.28) for both gastric and intestinal cells, respectively, suggesting that AFB1

142

CHAPTER III. RESULTS AND DISCUSSION

transport is preferable in the AB direction, and in the case of intestinal absorption the UR > 2 may suggest the presence of an uptake transporter in the apical membrane (402). In fact, intestinal transporters such as organic cation transporter (OCT) and organic anion transporters (OAT) were found to mediate the uptake of aflatoxin B1 (395, 396). The influence of transporting mycotoxins in mixture was also noticed in the uptake/efflux ratios of DON and AFB1. In the case of DON a similar efflux ratio (0.86) in Caco-2 cells was reported by Kadota and colleagues (242) when transported individually, but its simultaneous transport with other mycotoxins increased the UR of DON and decreased the ER, while decreased UR (0.45 and 1.74) of AFB1 and increased the ER (2.24 and 0.58). This may suggest that the co-transportation of mycotoxins might benefit the uptake of some mycotoxins to the detriment of others across the gastric and intestinal monolayers. The UR and ER ratios of DON were all lower than 2 showing that active transport was negligible under the experimental conditions and indicating that DON absorption/excretion is not modulated by P-glycoprotein (PgP) and multidrug resistance associated proteins (MRPs)

(243). On the other hand, AFB1 exhibited high UR (3.76 and 1.74) when transported isolated and in mixture across Caco-2 monolayers and a high ER (2.24) when transported in mixture across NCI-N87 monolayers, which may suggest the involvement of transporters as AFB1 was previously mentioned as substrate of uptake and efflux transporters. The changing of these ratios when simultaneously transported might suggests that the co-ingestion of multiple mycotoxins may influence their cell uptake and the ability of cells to excrete some xenobiotics. The use of Caco-2 cell model to predict the human fraction absorbed of pharmaceuticals has often been used by correlating the AB apparent permeabilities across Caco-2 monolayers of molecules with the experimental human fraction absorbed data, resulting in a sigmoidal relationship between the human fraction absorbed and the log (Papp) of molecules (330, 331). Thus, with this correlation, the in vitro permeability of a compound in Caco-2 cells can be used to predict the absorption in humans. The fraction absorbed (FA%) values are considered more intuitive than the Papp values for understanding intestinal absorption of compounds offering the additional benefit of assessing permeability ranking as sparingly (0-20%), intermediately (20-80%), and completely (80-100%) absorbed. Figures III.28A and Figure III.28B show the FA% of each mycotoxin isolated and in mixture as well as their position in the sigmoidal curve according to their percentage of absorption. AFB1 is the mycotoxin with the highest FA percentage (> 96%) both isolated and mixed transported; followed by DON with FA% of 72.8 and 82.9 when transported isolated and in mixture; and finally, OTA with significant difference (p<0.05) between isolated (11%) and mixed (66%) transport. In this sense, AFB1 is classified as completely absorbed either

143

CHAPTER III. RESULTS AND DISCUSSION

isolated or in mixture, DON is intermediately absorbed isolated and completely absorbed in mixture, and OTA has the greatest difference being sparingly absorbed isolated and intermediately absorbed in mixture (Figures III.28A and Figure III.28B).

Figure III.28 – Fraction absorbed at intestinal level (FA%) of DON, AFB1, and OTA, isolated and in mixture (A); as well as their position in the sigmoidal curve according to their FA% – completely (red), intermediately (orange), and sparingly (yellow) absorbed (B). The sigmoidal curve was built according to Skolnik and colleagues (330) and Tavelin and colleagues (331). Data is expressed as mean ± SD of 3 independent experiments (n=3). AFB1 – Aflatoxin B1; DON – Deoxynivalenol; and OTA – Ochratoxin A. Values with different letters differ significantly (p<0.05) from the given mean (lowercase letters for isolated transport, uppercase letters for transport in mixture. * significant difference (p<0.05) on FA% between mycotoxins transported isolated or in mixture.

A relevant finding of the present work shows that the fraction absorbed, calculated to predict the final % of intestinal uptake (330, 331), is significantly increased for OTA when exposure occurs in mixture (Figure III.28A). Berger and colleagues (239) and Schrickx and colleagues (391) proposed OTA as a substrate of efflux transporters (MRP2 and BCRP) decreasing its uptake in Caco-2 cells. However, the co-exposure of OTA with some polyphenols increased OTA transport across Caco-2 monolayers probably due to the competition for the MRP2 efflux pump (403). Moreover, the co-administration of OTA and DON in pigs resulted in the double of concentration of OTA in liver, muscles and kidney in comparison with OTA single administration (392, 404). This results corroborate the findings of the present study suggesting that both absorptive epithelia in the GIT show different absorptive patterns when mycotoxins are transported isolated or in mixture, which may be justified concerning that some mycotoxins, as AFB1 and OTA, share absorptive pathways (i.e. uptake/efflux transporters). However, further investigation on combined ingestion of toxins and their mixed transport should be included in such evaluations, concerning the frequent co-occurrence and consequent co-exposure of these mycotoxins. Moreover, this 144

CHAPTER III. RESULTS AND DISCUSSION data should be taken into consideration in future assessment of human toxicity by these compounds.

145

CHAPTER III. RESULTS AND DISCUSSION

III.3.5 TOXICOLOGICAL INTERACTIONS BETWEEN MYCOTOXINS

In this study four mycotoxins – AFB1, DON, OTA, and FB1 – from four different chemicals groups were selected to evaluate the individual and binary toxicity, in terms of additive, antagonistic and synergistic toxicity towards the intestinal Caco-2 cells and the hepatic HepG2 cells.

III.3.5.1 INDIVIDUAL CYTOTOXICITY OF MYCOTOXINS

Human-derived intestinal and hepatic cells (Caco-2 and HepG2) were selected as cellular models for the present study considering their human-derived origin, with the intent of providing more realistic situation concerning human health risk, and also considering previous reports describing toxicological effects of mycotoxins in these models in order to ease the comparison of results (265, 273, 274). The effect of all four mycotoxins on Caco- 2 and HepG2 cells viability (%) was assessed by the MTT assay after 72h incubation to indirectly determine their cytotoxicity by measuring the mitochondrial activity of viable cells. Three of four mycotoxins (AFB1, DON, and OTA) had a dose-effect response on both intestinal and hepatic cells, showing different dose-effect curves and consequently distinct

IC50 values for each compound. On the other hand, FB1 presented no toxicity, equally on both cell lines, at the concentrations used (0.625-20 µM) (Figure III.29).

146

CHAPTER III. RESULTS AND DISCUSSION

Figure III.29 – Percentage of viability (%) of proliferating Caco-2 and HepG2 cells after isolated exposure to serial dilutions of AFB1, DON, FB1 and OTA. Cytotoxicity was assessed by MTT assay. Data are mean ± SD of at least two independent experiments (8 replicates).

The parameters (r, m, Dm) of the dose-effect response curves of individual mycotoxins are presented in Table III.13. In terms of isolated toxicity, DON was the mycotoxin with the highest toxic effect on both cells viability, having the lowest IC50 values of 2.91 and 0.68 µM for Caco-2 and HepG2 cells, respectively. Based on individual IC50 values, the mycotoxins were ranked according to their cytotoxicity: DON (IC50=2.92 µM) > OTA (IC50=7.63 µM) >

AFB1 (IC50=19.3 µM) for exposure on Caco-2 cells; and DON (IC50=0.68 µM) > AFB1

(IC50=1.51µM) > OTA (IC50=21.2 µM) for exposure on HepG2 cells. Table III.13 shows the concentration of each mycotoxin needed to decrease cells viability in 10% (IC10), 25% (IC25), 50% (IC50), 75% (IC75) and 90% (IC90). The calculated IC50 values

147

CHAPTER III. RESULTS AND DISCUSSION

for AFB1, DON, and OTA values are similar to those reported in previous studies on both Caco-2 cells (257, 271, 405), and HepG2 cells (265, 406, 407).

Table III.13 – Isolated effect parameters calculated from Compusyn software and based on Chou-Talatay method for combination assays (332). (A) Caco-2 cells r m IC10 IC25 IC50 (Dm) IC75 IC90

AFB1 0.95 0.74 1.01 4.41 19.3 84.3 368 DON 0.92 0.97 0.31 0.94 2.92 9.02 27.9 FB1 - - n.f n.f n.f n.f n.f OTA 0.97 0.99 0.82 2.5 7.63 23.2 70.7 (B) HepG2 Cells r m IC10 IC25 IC50 (Dm) IC75 IC90

AFB1 0.98 1.4 0.31 0.67 1.51 3.3 7.24 DON 0.93 1.07 0.09 0.24 0.68 1.89 5.3 FB1 - - n.f n.f n.f n.f n.f OTA 0.91 3.03 10.2 14.7 21.2 30.4 43.7 r values represent the goodness of the fit-curve; and m signifies the shape of the dose-effect curve (m=1, hyperbolic, m>1 sigmoidal, m<1 flat sigmoidal). All inhibition concentrations (ICx) are expressed as µM; n.f – not found.

The dose-response shapes of AFB1 and DON exposure to HepG2 cells are in agreement with Zhou and colleagues (265). No other studies, using the same cell lines and Chou- Talatay theorem, were found to allow comparing dose-response shapes for the other mycotoxins and combinations. The non-toxic effects of FB1 are in accordance with Clarke and colleagues (405) for Caco-2 cells, and 200 µM was reported as IC50 value of FB1 on HepG2 (408), which validates the non-toxicity of FB1 found in our study, since our maximum concentration was 20 µM. Furthermore, FB1 showed also to be not toxic to immortalized BRL 3A rat liver cells (409) and to swine jejunal epithelial cells (410). DON exhibited the highest individual toxicity in both cells, which may result from trichothecenes’ multiple inhibitory effects on different targets of eukaryotic cells (e.g. protein inhibition, DNA and RNA synthesis, mitochondrial inhibition function, effects of cell division and membrane toxicity) (411). This specific mycotoxin has shown to induce apoptosis, dysfunction, and oxidative stress in the mouse kidney (264). DON’s high cytotoxic effect on HepG2 cells is in agreement with Zhou and colleagues (265). Moreover, tissues with high rates of cell turnover, which is the case of intestinal cells, are very susceptible to mycotoxins, due to their ability to strongly compromise the homeostasis of self-renewing capacities of cells causing the impairment of epithelial barrier increasing membrane permeability (274). Our study revealed that lower levels of DON than those found in several foodstuffs, 2.92 and 148

CHAPTER III. RESULTS AND DISCUSSION

0.676 µM, are able to reduce 50% the viability of Caco-2 and HepG2 cells, respectively. These results highlight the huge importance of controlling the concentrations of mycotoxins in the environment, and the need to infer the toxicity of DON in animals and humans, due to its high prevalence in food at concentrations that strongly reduce intestinal and liver cells viability. Moreover, researchers believe that DON may cause various chronic intestinal inflammatory diseases, such as inflammatory bowel disease, as well as contribute to food allergies, particularly in children (230). However, despite the recent studies on DON’s toxicity, evidence in animal models is scarce, which contributed to the IARC classification of DON as not classifiable as to its carcinogenicity to humans (group 3); and for this reason the toxicity of this molecule has not been re-evaluated since 1993. Considering all these recent data suggesting the high toxicity of DON in in vitro assays, it is needed to confirm such results using animal models. In the past few years, several studies reported the impact of feeding piglets with feed contaminated with DON (0.9-4 mg/kg feed) over 10 days to 11 weeks. This chronic exposure of DON induced oxidative stress, increased intestinal permeability, damaged villi, inhibited protein synthesis and cell proliferation (263, 412-417), contributed for histological lesions (263, 415, 418); and, reduced the number of goblet cells and lymphocytes, induced up regulation of cytokine expression in jejunum and ileum and down regulation of interleukin-1 beta (IL-1β) and IL-8 expression in ileum (413, 419).

III.3.5.2 CYTOTOXICITY OF BINARY COMBINATION OF MYCOTOXINS

Combined assays were performed using those mycotoxins that showed individual toxicity to infer if their combined effects were synergic, additive or antagonistic. The combination ratios were calculated on the basis of equipotent IC50 values obtained from the individual exposure experiments, so that the contribution to the effect of each mycotoxin would be equal (277) (Table II.1). That is the reason for the FB1 exclusion from combination experiments, as this mycotoxin did not show individual toxicity as referred (Figure III.29). All three combinations: AFB1-DON, AFB1-OTA, and OTA-DON presented dose-dependent effect in both cell lines, strongly decreasing cells viability (%) at higher concentration ratios (Figure III.30).

149

CHAPTER III. RESULTS AND DISCUSSION

Figure III.30 – Percentage of viability (%) of proliferating Caco-2 and HepG2 cells after 72h of binary combined exposure to AFB1-DON, AFB1-OTA, OTA-DON. Cytotoxicity was assessed by MTT assay. Data are mean ± SD of at least two independent experiments (8 replicates).

The realistic possibilities of co-occurrence and the concentrations at which mycotoxins are found in foodstuffs were also taken into consideration in the experimental design. The co-occurrence of AFB1, DON, FB1, and OTA has already been described in feedstuff and food (48, 401). For example, AFB1, DON and OTA were reported as the most common combination (29%) found in contaminated Spanish barley samples (420), with maximum concentrations found of 0.34 µg/kg (0.109 nM), 1111.3 µg/kg (3.75 µM), and 3.53 µg/kg (8.74 nM), respectively (421, 422). In this case, DON concentration is within the range tested in our study. Indeed, DON’s maximum concentration of 3.75 µM can be considered as realistic in the human gut, assuming that 0.5-7 µM are realistic values to be found at intestinal level (243). The concomitant occurrence of trichothecenes, aflatoxins and ochratoxins was verified in several human tissues (liver, lung and brain) at levels up to 18, 5 and 10 µg/kg (423) as well as in urine in the concentrations of 5.72, 9.4 and 14.6 µg/L, respectively (424). The mycotoxins used for binary combinations also impact the cells viability, causing different toxic effects depending on the combination. For example, OTA-DON was the most toxic combination for Caco-2 cells, while AFB1-DON was the combination with the highest 150

CHAPTER III. RESULTS AND DISCUSSION toxicity in HepG2 cells. As Figure III.30 shows, for Caco-2 cells, AFB1-OTA began as the less toxic combination at lower concentrations, increasing its toxicity at higher concentrations, thus becoming as toxic as AFB1-DON. In the case of HepG2 cells, the combinations can be ranked as AFB1-DON > OTA-AFB1 > OTA-DON in terms of toxicity at all levels of concentrations. The dose-response shapes of AFB1-DON combination exposure to HepG2 cells are in agreement with Zhou and colleagues (265). Although AFB1 and OTA may not be reported at higher concentrations than DON, our study shows that combining AFB1-DON, OTA-DON, and AFB1-OTA at low concentrations may indeed contribute for synergism (Table III.14) increasing the concern over the toxicity of mycotoxins co-occurrence. In most cases, combining mycotoxins required lower concentrations to reduce the cells viability at a certain level, when comparing with the mycotoxins alone

(Table III.13). For instance, the IC50 of the mixture AFB1-DON on Caco-2 cells was obtained with half of the concentration of individual exposure.

151

CHAPTER III. RESULTS AND DISCUSSION

152

CHAPTER III. RESULTS AND DISCUSSION

For HepG2, a 10-times reduced concentration was obtained for the same toxic level

(IC50) (Tables III.14 and Table III.15). Moreover, for those combinations that had synergic effect, dose reduction indices (DRI) were calculated, at low (IC25), medium (IC50), and high

(IC75) inhibition levels (Table III.15).

Table III.15 – Dose reduction index values (DRI) at inhibition concentrations of 25, 50 and 75% for binary combinations of AFB1, DON and OTA, in Caco-2 and HepG2 cells that showed to have synergistic effect.

(A) Caco-2 cells

Combination ratio IC25 IC50 IC75

AFB1 - 2.20 3.85 7:1 DON - 2.33 2.89

AFB1 - - 3.75 2:1 OTA - - 2.06

OTA 3.46 3.18 2.92 3:1 DON 3.91 3.65 3.41

(B) HepG2 cells

Combination ratio IC25 IC50 IC75

AFB1 14.3 11.0 8.50 2:1 DON 10.0 9.88 9.75

AFB1 1.91 2.32 2.83 1:14 OTA 2.92 2.33 1.86

OTA - - - 31:1 DON - - -

For Caco-2 cells, the DRI values ranged from 2.20 to 3.85, 2.06 to 3.75, and 2.92 to 3.91 for AFB1-DON, AFB1-OTA and OTA-DON, respectively; whereas for HepG2, the DRI values ranged from 8.5 to 14.2 and 1.85 to 2.92 for AFB1-DON and AFB1-OTA, respectively. Thus, these values are within the same range (between 2 and 4) on Caco-2 cells, despite the mycotoxin combination, but on HepG2 cells, DRI values about 10-times greater were found for AFB1-DON when compared with AFB1-OTA. This shows that although having both synergic effects at the selected inhibition concentrations, the combinations may have distinct synergic strength. Moreover, from the calculated combination index (CI) values that quantify the effect of mycotoxins’ combination (synergism if CI<0.9, additive if 0.9

153

CHAPTER III. RESULTS AND DISCUSSION

CI>1.1), effect-oriented graphs (CI-plots) and dose-oriented graphs (isobolograms) were built for all combinations on both cells (Figures III.31 and Figure III.32). The intestinal cells showed signs of antagonistic effect at lower levels of toxicity when exposed to AFB1-DON and AFB1-OTA mixtures, and synergism at higher toxic levels; whilst DON-OTA combination showed synergism at all inhibition levels (Figure III.31A). The Figure III.31B shows the isobolograms of each combination, at three inhibition concentrations (IC25, IC50, and IC75), for Caco-2 cells. These graphs are a two dimensional representation of a dose- oriented effect, where each axis represents one mycotoxin used in the mixture. For each inhibition concentration (ICx), a diagonal line connects the ICx of each toxin alone, and a dot represents the ICx of the combination. The dots position determines if the combination is antagonistic (above the line) or synergistic (below the line), and the distance between the dot and the line determines the strength of the combined toxic effect. The toxicity of combined mycotoxins differed with the inhibition concentration and the selection of mycotoxins for combination, e.g. AFB1-DON had antagonistic effect at IC25 and synergic effect at IC50 and IC75 for Caco-2 cells. Gastrointestinal tract represents the first barrier met by mycotoxins following the gastrointestinal route, and our study shows that proliferating Caco-2 cells suffer antagonistic and synergistic effects after exposure (72h) to binary combined mycotoxins (Table III.14 and Figure III.31). Combinations of AFB1 with DON and OTA showed antagonism: AFB1-

DON only at lower inhibitions levels (IC10 and IC25); while AFB1-OTA had antagonistic effect on Caco-2 cells until IC50. Such AFB1-OTA antagonistic effects on intestinal cells may be justified with OTA strongly antagonistic behaviour towards AFB1, because of OTA’s enhanced expression of metabolic action by CYP3A4 and attenuation of AFB1-DNA adduct formation lowering the AFB1 toxicity (255). The synergism observed on Caco-2 cells with DON-OTA combination was also reported by Cano-Sancho and colleagues (271). OTA mechanisms of action are not clearly determined, but apparently OTA may disrupt phenylalanine metabolism, reduce gluconeogenesis and induce apoptosis via protein/DNA synthesis inhibition (425). Therefore, the co-occurrence of DON and OTA may become a real hazardous problem at intestinal level, since both mycotoxins are referred as promoters of impairment of intestinal barrier function and increase membrane permeability. The capacity of mycotoxins to contribute to the oxidative stress was pointed out as a possible contribution for synergism effects on Caco-2 cells after combined mycotoxins exposure (257, 425).

154

CHAPTER III. RESULTS AND DISCUSSION

Figure III.31 – Mycotoxins combination plots based on Chou and Talatay combination index theorem after 72 h exposure on Caco-2 cells: (A) combination index plot (CI-Plot) for binary combination of mycotoxins: AFB1- DON, AFB1-OTA, and DON-OTA. The CI values were calculated from data obtained from at least two independent experiments on the basis of equipotent mycotoxins combinations; the vertical bars indicate 95% confidence intervals for CI values based on sequential dilution analysis (Chou 2006). CI <1, CI=1, and CI >1 indicate synergistic, additive, and antagonistic effects, respectively. Dashed lines indicate the additive level, which separates synergism and antagonism sides; (B) Classic isobolograms illustrating the combined toxicity of

AFB1-DON, AFB1-OTA, and OTA-DON at IC25, IC50 and IC75. The points are the mean concentration of dose- response MTT cytotoxicity curves for each mycotoxin or their binary combination. All these results were calculated using Compusyn software.

The combined exposure of mycotoxins to liver cells was equally assessed. This resulted in a strong synergism for AFB1-DON at all inhibition levels; a slight synergism or nearly additive effect for AFB1-OTA combination; and moderate antagonism at lower inhibitions levels for OTA-DON combinations, and nearly additive at higher inhibition levels (Figure III.32A). Regarding the isobolograms for HepG2 cells (Figure III.32B), the toxicological 155

CHAPTER III. RESULTS AND DISCUSSION

interactions relied more on the inhibition concentrations tested, with AFB1-DON showing strong synergistic effects at all inhibition levels, AFB1-OTA having a synergism effect at all concentrations, while DON-OTA combination showed an antagonistic effect at IC25, moderate antagonism at IC50, and an additive effect at IC75.

Figure III.32 – Mycotoxins combination plots based on Chou and Talatay combination index theorem after 72 h exposure on HepG2 cells: (A) combination index plot (CI-Plot) for binary combination of mycotoxins: AFB1- DON, AFB1-OTA, and DON-OTA. The CI values were calculated from data obtained form at least two independent experiments on the basis of equipotent mycotoxins combinations; the vertical bars indicate 95% confidence intervals for CI values based on sequential dilution analysis (Chou 2006). CI <0.9, 0.91.1 indicate synergistic, additive, and antagonistic effects, respectively. Dashed lines indicate the additive level, which separates synergism and antagonism sides; (B) Classic isobolograms illustrating the combined toxicity of

156

CHAPTER III. RESULTS AND DISCUSSION

HepG2 cells are described as the best model of liver cells to reflect the metabolism of xenobiotics, as these cells express intact and inducible phase I and II enzymes, which play a crucial role in activation and detoxification of xenobiotics (426, 427). Similarly to Caco-2 cells, the combination of mycotoxins caused different toxic effects on HepG2 cells. An expected strong synergism was observed for AFB1-DON combination at all inhibition levels, regarding the high individual toxicities of DON and AFB1 on these cells, allied with the hepatotoxic effects of aflatoxins already well described in literature (428). Furthermore, DON-induced DNA damage in HepG2 cells has been linked to oxidative stress (427). In the liver, AFB1 is biotransformed by cytochrome P450 to its highly reactive metabolite – AFB1- 8,9-epoxide – which binds to nucleic acids forming DNA adducts (429). Synergistic effects of AFB1-DON were also reported on HepG2 cells, and other liver cell models (BRL 3A rat liver cells) (265, 409). Though, the toxic effects of AFB1-DON were found additive on primary hepatocytes of Cyprinus carpi (430). Considering the strong synergism observed in this combination (IC50AFB1+DON = 0.14+0.07 µM), meaning a 10-times concentration reduction compared with mycotoxins alone to achieve the same toxicity level, their co- occurrence is a human health risk concern. Combining DON with 15-acetyl-deoxynivalenol, nivalenol, or fusarenon-X has also resulted in synergistic effects in human gastric epithelial (GES-1) cells (196). Our study allied with recent literature data reporting the high toxic effect of DON alone and combined with other mycotoxins suggests that even though DON has not been classified as to its carcinogenicity to humans, this mycotoxin may present a serious threat to humans’ health, mainly when co-occurring with other mycotoxins contributing to synergism. Thus, to study the toxic effects of DON in combination with AFB1 and OTA, as well as with other mycotoxins, using in vivo animal models shall be subject for next investigations. Combination of AFB1-OTA appear to have nearly additive/slight synergism effects when exposed to HepG2 cells, as OTA reduces AFB1 induced DNA damage in HepG2 cells, which agrees with the additive cytotoxic effect reported by Corcuera and colleagues (431) for the same cell model. To the best of our knowledge, no other studies reported the combined effect of OTA- DON on HepG2 cells or other liver cells. Our study shows antagonism and nearly additive effects at lower and higher inhibition levels, respectively. Despite the described oxidative stress contribution of DON on HepG2 cells toxicity, the antagonism effect observed may be due to the described strong antagonistic behaviour of OTA toward other mycotoxins (255, 431). Also, even though DON and OTA may have distinct metabolizing pathways, they share common mechanisms of action, such as contribution for oxidative stress by ROS formation in HepG2 cells, as well as both contributing for DNA damage (427, 432). Thus,

157

CHAPTER III. RESULTS AND DISCUSSION

more research regarding the metabolism of mycotoxins and their interaction is needed to confirm such antagonism effects in liver cells. Our study did not approach the molecular mechanisms underlying mycotoxins’ interactions, though, considering the results obtained, information about the mechanistic pathways of mycotoxins mixtures at intestinal and hepatic level is needed to provide data for the establishment of updated and more realistic risk assessment strategies considering their toxicological effects on human health.

III.3.6. FINAL REMARKS

III.3.6.1 IMPACT OF COOKING AND IN VITRO DIGESTION

Mycotoxins spiked in chicken breast muscle are differently degraded after microwaving and oven cooking. Else, the use of roasting bags while cooking does not add any advantages in comparison to microwaving or oven cooking itself in what respects mycotoxins mitigation. However, adding herbs prior to cooking has a strong impact on mycotoxins reduction when compared with cooking without using herbs. In this sense, seasoning meat with herbs prior to cooking could be indicated as a mitigation strategy to reduce mycotoxins content in contaminated meats. Notwithstanding one should be always aware that herbs can also be naturally contaminated with mycotoxins. Concerning mycotoxins bioaccessibility after different cooking methods, adding herbs positively correlates with the increase of AFB1, AFG1, and AFG2 bioaccessibility, in comparison with not using during cooking. This suggests that even though mycotoxins can be in lower amounts after cooking with herbs, these may be more bioaccessible after in vitro digestion. In general, cooking itself is effective to reduce mycotoxins content in chicken meat, but adding herbs prior to cooking additionally reduces their content up to 60%. However, while adding herbs seems to beneficiate the reduction of mycotoxins during cooking, it appears to increase their bioaccessibility after in vitro digestion.

III.3.6.2 TRANSPORT ACROSS GASTROINTESTINAL EPITHELIUM

In this study, it was evaluated for the first time the transport of four prevalent mycotoxins across human differentiated NCI-N87 gastric cells, isolated or in mixture, in parallel with transport studies in the well-known Caco-2 intestinal model. At the best of our knowledge the intestinal transport of a mixture of mycotoxins across Caco-2 has not been previously explored. 158

CHAPTER III. RESULTS AND DISCUSSION

This study showed that AFB1 is rapidly transported across gastric and intestinal cells, in both directions, followed by DON; while OTA and FB1 have a selective transport only occurring from the apical to basolateral side, in both cells for OTA, and only at gastric levels for FB1, probably due to their pH dependence (preferably transported in acidic sides of monolayers). This data should be taken into consideration in future assessment of human exposure to these compounds. Moreover, the co-exposure of mycotoxins strongly influenced their cell uptake/efflux across gastric and intestinal monolayers significantly increasing OTA and DON intestinal fraction absorption (FA %) changing from sparingly to intermediately absorbed (OTA) and from intermediately to completely absorbed (DON). This results suggest different absorptive patterns in both epithelia in the gastrointestinal tract when mycotoxins are transported isolated or in mixture, which may be justified concerning some of their shared absorptive pathways as AFB1 and OTA that use the same uptake/efflux transporters. However, further investigation on combined ingestion of toxins and their mixed transport effects should be included in future evaluations considering the unavoidable and frequent co-occurrence and consequent human co-exposure to mycotoxins, thus, this studies configure an open field for research. Moreover, this data will be necessary to be taken into consideration in future assessment of human toxicity by these compounds.

III.3.6.3 INTESTINAL AND HEPATIC TOXICOLOGICAL INTERACTIONS

Mycotoxins interactions were evaluated using two different model systems appropriate for evaluation of intestinal or liver toxicity and an experimental design that included realistic doses of mycotoxin. Caco-2 and HepG2 cells were more sensitive to DON alone than to other mycotoxins. Additionally, when combined, OTA-DON and AFB1-DON were the most toxic combinations for Caco-2 and HepG2, respectively, having both synergistic effects at all inhibition levels. Therefore, even though DON has not been classified as to its carcinogenicity to humans, this mycotoxin may present a serious threat to human health. The referred synergistic effects are of biological importance since these mycotoxins have proved to co-occur in the environment and these results suggest that the human exposure to DON throughout any route must be of huge concern. Further investigation as well as mechanistic studies are needed to deeper infer the actual toxic effects of these combinations.

159

CHAPTER IV – GENERAL DISCUSSION

Mitigation strategies are mandatory actions to keep chicken meat safe and maintain its nutritionally value. The main goal of this thesis was based on this idea and tried to understand the feasibility of culinary practices to prevent/reduce the presence of chemicals hazards in chicken meat upon cooking as well as after digestion. Microwave (radiation) or oven cooking (dry heat) did not differ on the formation of protein/lipid oxidation products neither concerning mycotoxins reduction, despite their different heating systems. Nervetheless, in the case of ACDs, oven cooking seemed to benefit their reduction. Both cooking methods promoted nutrients oxidation, contributing to the loss of PUFAs and free amino acids while forming concerned aldehydes, such as HNE and MDA that may cause several health concerns. Keeping that in mind, ACDs reductions were expected after cooking, as some of them are degraded by oxidative process, but only AMX, CTC, and TYL were highly reduced (>50%), partially justified by molecular rearrangement (isomerization/epimerization) and dechlorination reactions. In addition, the amount of ACDs content after cooking relies on more than one variable, like the initial level of contamination, type of cooking, or ACD type, meaning that the cooking stability of one ACD class cannot be predicted by studying only one family-related compound (e.g. TCs family). Concerning mycotoxins, these were all reduced after cooking, except T2. Their stability to cooking is considered structurally-related as mycotoxins grouped by chemical family in the agglomerative hierarchical clustering analysis. Overall, despite the loss of nutritional value of meat concerning protein and lipid content, cooking itself could be used as a mitigation strategy to reduce some ACDs and mycotoxins. Besides the ability of cooking to reduce some contamination in chicken meat, this thesis also evaluated the inclusion of natural ingredients as mitigation strategies, like the herbs and beer, widely used in European countries to give flavour to food, and the roasting bags, used to keep the moisture during cooking. The type of ingredients – oregano/beer or roasting bag/mixture of herbs – differently affected the behaviour of chemical hazards. This thesis tested lower oregano contents than those found in literature, to get as close as a real situation, demonstrating a strong preventive behaviour against the formation of the three aldehydes and keeping the nutritional value of meat, i.e. higher total and ω-6 PUFAs contents in comparison with cooked samples without ingredients. In the same way, adding herbs reduced the content of almost all mycotoxins after cooking up to 60%. Herbs addition was classified as a variable important for the projection in the PLS regression analysis for AFB1, AFG1 and AFG2, suggesting a negative correlation with the percentage of mycotoxins after cooking. On the contrary, adding herbs hindered the reduction of most of ACDs. This behaviour agrees with

163

CHAPTER IV – GENERAL DISCUSSION

literature for ACDs, as the case of SAs and FQs that are degraded by oxidative processes, since the increment of antioxidant components would prevent the oxidative processes and protect ACDs from degradation. The opposite behaviours observed may result from the complexity of the matrices studied (meat and oregano) that contain both antioxidant and pro-oxidant compounds and could benefit the reduction of some ACDs. In the case of DNC, the preventive behaviour of herbs against degradation avoids the release of p-NA, a compound of concern in cooked chicken meat, but it raises a dual problem since adding herbs prevents human exposure to p-NA, but it does not help reducing DNC exposure to humans. Thus, new strategies should be studied to mitigate DNC in chicken meat avoiding the formation of concerned compounds like p-NA. Despite beer being referred as a rich source of polyphenols and used as a mitigation strategy to prevent the formation of harmful compounds in cooked meat, herein the addition of beer did not prevent meat from oxidation equally forming MDA, HNE, HEX, and shiff bases as cooking themselves. Moreover, higher losses on ω-6 PUFAs and ω-3 PUFAs were observed. In addition, oven cooking with beer was the only cooking method whose carbonyls content increased, and whose free amino acids contents lowered in comparison with those cooked with oregano. The alcohol of beer (5%v/v), as well as the presence of other reactive aldehydes not measured in our experiments are plausible causes of the non- protective effects against oxidation. Beer had no effect or a similar effect as observed with herbs during cooking of contaminated meat with ACDs, avoiding the reduction of some ACDs or reducing the contents of others. For mycotoxins, the addition of beer was not tested, instead the use of roasting bags was evaluated, which proved to have no influence on their stability. Overall, herbs are the best mitigation strategy to reduce oxidation of meat and mycotoxins content in case of contamination, and preserve PUFAs. However, these may not help reduce ACDs content in a realistic situation. In this sense, other mitigation strategies should be tested to efficiently reduce ACDs from chicken meat concerning that cooking themselves and the addition of extra ingredients (herbs or beer) proved to be quite inefficient in reducing their contents. All these variables indicate that residue data on raw muscles may not reflect the real exposure of consumers to ACDs or mycotoxins, or even MDA and HNE. In this sense, relying only on meat contaminants levels may not be the best approach to estimate exposure to these hazards and therefore calculate acceptable daily intakes because cooking as well as some other culinary practices showed to affect their contents. The release of nutrients and/or hazards during in vitro digestion of cooked meat revealed to be more dependent on the presence or not of ingredients rather than the type of cooking.

164

CHAPTER IV – GENERAL DISCUSSION

Lipid and protein oxidation increased during digestion, with increased content of MDA, carbonylation and Schiff bases, whereas HNE and HEX decreased, probably due to their conversion into tertiary oxidation products (volatile compounds) or interaction with protein residues. All this oxidation likely compromised the proteolysis and consequent protein quality of chicken meat. Concerning ACDs, the determination of their bioaccessibility at the realistic concentration (0.1 mg/kg) was not possible, although the samples fortified at 1 mg/kg exhibited bioaccessibility up to 60% for all ACDs, except AMX and DNC that were not detected. In the same way, mycotoxins also had lower bioaccessibility than those found in literature regarding in vitro digestion of cereal-based foods. The low bioaccessibility of ACDs and mycotoxins likely result from interaction with the matrix, such as drug-food interactions with cations, proteins and bile salts from matrix, or protein interaction with mycotoxins, hindering their quantification by LC-QqQ-MS/MS. These food-hazard interactions during in vitro digestion introduce some instrumental limitations to assess final intake of these compounds, thus ACDs and mycotoxins determination in digested meat matrices should not be based only on considering their free form. The addition of herbs or beer during cooking also influenced the fate of nutrients and hazards during digestion. The higher PUFAs content observed in digested samples with added ingredients (herbs or beer) suggests that these ingredients are facilitators of fatty acids digestion or preventers of further lipid oxidation during digestion. Moreover, digested samples with oregano exhibited the highest content of ω-3 and ω-6 PUFAs, even though HEX and HNE increased after digestion allied with the loss of antioxidant ability of the herb against the oxidant agents during the digestive process. Unlike ACDs content after cooking, the addition of herbs and/or beer was considered a variable with significance concerning ACDs bioaccessibility. Most ACDs exhibited higher bioaccessibility in cooked samples with herb or combination of herbs and beer, although this behaviour was not observed for all ACDs. In the same way, digested cooked meat with herbs exhibited higher mycotoxins bioaccessibility. Moreover, a PLS regression classified the herbs as variables important for the projection of AFB1, AFG1 and AFG2, suggesting a positive correlation with the percentage of mycotoxin bioaccessibility. In this sense, the herbs represent the best mitigation strategy to reduce the exposure of humans to lipid and protein oxidation products after ingestion of cooked meat. On the other hand, the beneficial impact on the other studied hazards is not so clear, since ACDs were not significantly reduced by herbs during cooking and herbs increased their bioaccessibility, and in the case of mycotoxins although herbs reduce their content after cooking, an increase of the bioaccessibility was observed on the remaining mycotoxins. However, even though being highly bioaccessible, as cooking with herbs highly reduced their contents, the

165

CHAPTER IV – GENERAL DISCUSSION

exposure to these mycotoxins after digestion is lower in samples with added herbs. Thus, herbs could be successfully applied as mitigation strategies to avoid lipid and protein oxidation during cooking and digestion preserving the nutritional quality of chicken meat, but their impact is not clear concerning ACDs and mycotoxins. The health concerns around mycotoxins toxicity and their co-occurrence in food, meat or others, has led to the second goal of this thesis which focused on evaluating the isolated and mixed absorption of mycotoxins across GIT epithelium and toxicity in target organs. The absorptive patterns of mycotoxins differed in both GIT epithelia depending on their isolated and mixture transport. This phenomenon happened because some mycotoxins share the same absorptive pathways, as are the case of AFB1 and OTA who share the same uptake/efflux transporters. These are prevalent mycotoxins who often contaminate food and feed, and the climate changes will likely result in outbreaks in food and feed. Thus, further investigation on combined ingestion of toxins and their mixed transport should be taken into consideration in future assessment of human toxicity by these compounds concerning the frequent co-occurrence and consequent co-exposure to humans. Among the mycotoxins tested, DON exerted the highest toxic effects in intestinal and hepatic cells. This raises a concern since this mycotoxin has not been classified as to its carcinogenicity to humans belonging to Group 3 in IARC evaluations since 1993. Based on our data and allied with literature reports, this mycotoxin may present a serious threat to human health. Binary combinations of mycotoxins caused different toxic effects regarding the mycotoxins used, level of inhibition and cell lines evaluated. OTA-DON and AFB1-DON were the most toxic combinations in intestinal and hepatic cells, respectively, having both synergistic effects at all inhibition levels. This evidences the high toxic effect of DON alone and combined with other mycotoxins and the serious health effects that may arise from DON exposure, mainly when co-occurring with other mycotoxins contributing to synergism. Thus, study the toxic effects of DON in combination with AFB1 and OTA on other target tissues, as well as with other mycotoxins, shall be subject for further investigation. In conclusion, the studies performed in this thesis provided novel information concerning the oxidation of nutrients during cooking and digestion, the impact of cooking practices on exposure to meat potential contaminants, such as ACDs and mycotoxins; the behaviour of the chemical hazards – cooking-induced, ACDs, and mycotoxins – in the presence of food ingredients (herbs, beer, or roasting bags), and the food-hazards interactions during in vitro digestion which may limit the precise assessment of the final intake of some of these compounds. Secondly, this thesis provided knowledge on interactions of four prevalent mycotoxins during gastric and intestinal absorption, as well as information on toxicological interactions on binary combinations of the same mycotoxins at intestinal and hepatic levels.

166

CHAPTER V – CONCLUSIONS AND FUTURE

PROSPECTS

CHAPTER V – CONCLUSIONS AND FUTURE PROSPECTS

The present thesis brings new information concerning two major issues: 1) the fate of three classes of chemicals hazards upon cooking and digestion; and 2) the influence of mycotoxins co-exposure concerning their gastrointestinal transport and toxicity. Concerning the first achievement, the different chemical hazards had different behaviours depending on mitigation strategies applied which conducted to different conclusions concerning their fates and best mitigation options to reduce their content. When dealing with a chicken meat portion free from hazards (e.g. ACDs or mycotoxins), oregano shall be added to avoid protein and lipid oxidation and consequent formation of hazardous compounds. This natural mitigation strategy will preserve PUFAs content, avoid loss of nutritional value of meat and diminish exposure to HNE and MDA after ingestion. This thesis only evaluated microwaving and oven cooking as cooking methods, and both seemed to equally influence nutrients oxidation, however, in the future other cooking methods should be investigated, as well as other herbs and/or other antioxidant ingredients should be explored as new mitigation strategies to prevent protein and lipid from oxidation. Moreover, as HNE and MDA are formed during cooking at a considerable level, further investigation should be done concerning their transport across gastric and intestinal monolayers, as well as their toxicological interactions. In the presence of a chicken meat contaminated with ACDs or mycotoxins the addition of ingredients, such as herbs, beer or the use of roasting bags does not bring considerable benefits, because their addition does not help reducing ACDs content, and in the case of mycotoxins, despite their reduction during cooking, the herbs benefit their bioaccessibility after digestion. In this sense, other mitigation strategies should be searched to properly reduce ACDs and mycotoxins in meat considering that cooking does not sufficiently reduce their content and adding ingredients has no impact or negatively impacts their reduction. Taken together, if withdrawal times are respected and if livestock animals are not fed with contaminated feedstuff, the chicken meat muscle will probably reach consumers in a hazards-free way, meaning that oregano herbs can be added to meat to flavour food and keep nutritional quality of meat. Concerning the second issue of this thesis, mycotoxins interact with each other either during absorption in epithelium and in terms of combined toxicity through synergism effects. OTA transport across Caco-2 monolayers exhibited a significant increase in the presence of mycotoxins, from sparingly absorbed (< 20%) to intermediately absorbed (~60%). Moreover, DON revealed to be highly toxic alone or in combination, which is of huge concern because this mycotoxin has not been classified as to its carcinogenicity to humans. This data reveals that mycotoxins co-occurrence is a real problem, and that this information should be taken into consideration in future evaluations. Thus, the combined exposure of

169

CHAPTER V – CONCLUSIONS AND FUTURE PROSPECTS

mycotoxins and/or other chemical hazards should be the subject of future studies that simulate chronic exposures at low doses. The outcomes accomplished are expected to represent a major input for the assurance of safety and nutritional relevance of cooked chicken meat for consumers, by combining data on the “natural” mitigation strategies and stability of several chemicals hazards during cooking, their behaviour through in vitro digestion; and in the case of mycotoxins, gastric and intestinal absorption, hepatic and intestinal toxicity patterns in a co-exposure situation.

170

REFERENCES

1. Wood JD. Chapter 20 - Meat Composition and Nutritional Value. In: Toldra´ F, editor. Lawrie´s Meat Science (Eighth Edition): Woodhead Publishing; 2017. p. 635-59. 2. Abete I, Romaguera D, Vieira AR, Lopez de Munain A, Norat T. Association between total, processed, red and white meat consumption and all-cause, CVD and IHD mortality: a meta-analysis of cohort studies. British Journal of Nutrition. 2014;112(5):762-75. 3. Menezes EA, Oliveira AF, França CJ, Souza GB, Nogueira ARA. Bioaccessibility of Ca, Cu, Fe, Mg, Zn, and crude protein in beef, pork and chicken after thermal processing. Food Chemistry. 2018;240:75-83. 4. Bushkin-Bedient S, Carpenter DO. Benefits versus Risks associated with consumption of fish and other seafood. Rev Environ Health. 2010;25(3):161-91. 5. Marin S, Ramos AJ, Cano-Sancho G, Sanchis V. Mycotoxins: Occurrence, toxicology, and exposure assessment. Food and Chemical Toxicology. 2013;60:218-37. 6. Pleadin, Persi N, Kovacevic D, Vulic A, Frece J, Markov K. Ochratoxin A reduction in meat sausages using processing methods practiced in households. Food additives & contaminants Part B, Surveillance. 2014;7(4):239-46. 7. Pleadin, Staver MM, Vahčić N, Kovačević D, Milone S, Saftić L, et al. Survey of aflatoxin B1 and ochratoxin A occurrence in traditional meat products coming from Croatian households and markets. Food Control. 2015;52:71-7. 8. Popp J, Pető K, Nagy J. Pesticide productivity and food security. A review. Agronomy for Sustainable Development. 2013;33(1):243-55. 9. Viegas O, Novo P, Pinto E, Pinho O, Ferreira IM. Effect of charcoal types and grilling conditions on formation of heterocyclic aromatic amines (HAs) and polycyclic aromatic hydrocarbons (PAHs) in grilled muscle foods. Food and chemical toxicology : an international journal published for the British Industrial Biological Research Association. 2012;50(6):2128-34. 10. Badry NEL. Effect of Household Cooking Methods on some Food Additives on Polycyclic Aromatic Hydrocarbons (PAHs) Formation in Chicken Meat. World Appl Sci J. 2010;9(9):963-74. 11. Khan MI, Min J-S, Lee S-O, Yim DG, Seol K-H, Lee M, et al. Cooking, storage, and reheating effect on the formation of cholesterol oxidation products in processed meat products. Lipids Health Dis. 2015;14:89. 12. Iwasaki M, Kataoka H, Ishihara J, Takachi R, Hamada GS, Sharma S, et al. Heterocyclic amines content of meat and fish cooked by Brazilian methods. J Food Compost Anal. 2010;23(1):61-9. 13. Jiang J, Xiong YL. Natural antioxidants as food and feed additives to promote health benefits and quality of meat products: A review. Meat science. 2016;120:107-17. 14. Persson E, Sjöholm I, Skog K. Heat and mass transfer in chicken breasts – effect on PhIP formation. European Food Research and Technology. 2002;214(6):455-9. 15. Ferraro JV, Plummer TW, Pobiner BL, Oliver JS, Bishop LC, Braun DR, et al. Earliest Archaeological Evidence of Persistent Hominin Carnivory. PloS one. 2013;8(4):e62174. 16. Smil V. Eating meat. Evolution, patterns, and consequences. Population and Development Review. 2002;28:599-639. 17. Ritchie H, Roser M. Meat and dairy production. Our world in data. Available: https://ourworldindata.org/meat-production#global-meat-production (last acessed 09.12.2019). 2019 [ 18. EC. European Comission REG (EC) No835/2004. 2002 Regulation (2004). EUOJ L139/55. 2004. 19. FAOSTAT. FAOSTATISTIC. Data on Food Supply - Livestock and Fish Primary Equivalent (http://www.fao.org/faostat/en/#data/QL) (accessed 2020 May 8th). 20. IARC. Consumption of red meat and processed meat. Lyon; 6-13 september, 2015. IARC Monogr Eval Carcinog Risks Hum. Lyon; 6-13 september, 2015: IARC monograph working gourp members; 2015.

172

REFERENCES

21. Cabrera MC, Ramos A, Saadoun A, Brito G. Selenium, copper, zinc, iron and manganese content of seven meat cuts from Hereford and Braford steers fed pasture in Uruguay. Meat science. 2010;84(3):518-28. 22. FAO. Food and Agriculture Organization of the United Nations). Meat consumption. http://www.fao.org/ag/againfo/themes/en/meat/background.html Accessed 07.12.19. 2014. 23. Lombardi-Boccia G, Lanzi S, Aguzzi A. Aspects of meat quality: trace elements and B vitamins in raw and cooked meats. Journal of Food Composition and Analysis. 2005;18(1):39-46. 24. INSA. Instituto Nacional de Saúde Dr Ricardo Jorge (2019). Tabela de Composição de Alimentos. Lisbon. 2019. 25. Pereira PMdCC, Vicente AFdRB. Meat nutritional composition and nutritive role in the human diet. Meat science. 2013;93(3):586-92. 26. Council NR. National Research Council (US) Subcommittee on the Tenth Edition of the Recommended Dietary Allowances. Recommended Dietary Allowances: 10th Edition. Washington (DC): National Academies Press (US); 1989. 6, Protein and Amino Acids. Available from: https://www.ncbi.nlm.nih.gov/books/NBK234922/. 1989. 27. Riccioni G, D'Orazio N, Menna V, De Lorenzo A. Fat Soluble Vitamins and Immune System: An Overview. European Journal of Inflammation. 2003;1(2):59-64. 28. Scollan N, Hocquette JF, Nuernberg K, Dannenberger D, Richardson I, Moloney A. Innovations in beef production systems that enhance the nutritional and health value of beef lipids and their relationship with meat quality. Meat science. 2006;74(1):17- 33. 29. Ricciotti E, FitzGerald GA. Prostaglandins and inflammation. Arterioscler Thromb Vasc Biol. 2011;31(5):986-1000. 30. Howe P, Meyer B, Record S, Baghurst K. Dietary intake of long-chain omega-3 polyunsaturated fatty acids: contribution of meat sources. Nutrition. 2006;22(1):47-53. 31. Laskowski W, Górska-Warsewicz H, Kulykovets O. Meat, Meat Products and Seafood as Sources of Energy and Nutrients in the Average Polish Diet. Nutrients. 2018;10(10):1412. 32. Zhang T, Li B. Occurrence, Transformation, and Fate of Antibiotics in Municipal Wastewater Treatment Plants. Critical Reviews in Environmental Science and Technology. 2011;41(11):951-98. 33. Ma, J, Zhai, G. Antibiotic Contamination: A Global Environment Issue. J Bioremed Biodeg 2014;5:e157. 34. Dmitrienko SG, Kochuk EV, Apyari VV, Tolmacheva VV, Zolotov YA. Recent advances in sample preparation techniques and methods of sulfonamides detection - A review. Anal Chim Acta. 2014;850:6-25. 35. Heshmati A, Kamkar A, Salaramoli J, Hassan J, Jahed GH. The effect of two methods cooking of boil-ing and microwave on tylosin residue in chicken meat. Iranian Journal of Nutrition Sciences and Food Technology. 2013;8:61-71. 36. Nisha AR. Antibiotic residues-a global health hazard. Veterinary World. 2008;1:375- 7. 37. EC. European Commission (EC); Commission Regulation (EU). No. 37/2010 of 22 December 2009 on pharmacologically active substances and their classiﬁcation regarding maximum residue limits in foodstuffs of animal origin. Off. J. Eur. Commun. 2010, 15, 3–36. 2010. 38. Omotoso AB, Omojola AB. Fluoroquinolone residues in raw meat from open markets in Ibadan, Southwest, Nigeria. . International Journal of Health, Animal Science and Food Safety. 2015;2(1):32-40. 39. Pena A, Silva LJ, Pereira A, Meisel L, Lino CM. Determination of fluoroquinolone residues in poultry muscle in Portugal. Analytical and bioanalytical chemistry. 2010;397(6):2615-21.

173

REFERENCES

40. Ahmadi F, Shahbazi Y, Karami N. Determination of tetracyclines in meat using two phases freezing extraction method and HPLC-DAD. Food Anal Methods. 2015;8(7):1883-91. 41. Olatoye IO, Ogundipe GAT. Quantitative analysis of oxytetracycline residue in beef and chicken meat from cities of southwest Nigeria. Bull Anim Health Prod Afr. 2013:39-48. 42. Salehzadeh F, Madani R, Salehzadeh A, Rokni N, Golchinefar F. Oxytetracycline residue in chicken tissues from Tehran slaughterhouses in Iran. . Pakistan J Nut. 2006;5:377-81. 43. Yamaguchi T, Okihashi M, Harada K, Konishi Y, Uchida K, Do MH, et al. Antibiotic residue monitoring results for pork, chicken, and beef samples in Vietnam in 2012- 2013. Journal of agricultural and food chemistry. 2015;63(21):5141-5. 44. Cheong CK, Hajeb P, Jinap S, Ismail-Fitry MR. Sulfonamides determination in chicken meat products from Malaysia. International Food Research Journal. 2010;17:885-92. 45. Diana Di Mavungu J, Monbaliu S, Scippo ML, Maghuin-Rogister G, Schneider YJ, Larondelle Y, et al. LC-MS/MS multi-analyte method for mycotoxin determination in food supplements. Food additives & contaminants Part A, Chemistry, analysis, control, exposure & risk assessment. 2009;26(6):885-95. 46. Hussein HS, Brasel JM. Toxicity, metabolism, and impact of mycotoxins on humans and animals. Toxicology. 2001;167(2):101-34. 47. Martins C, Assunção R, Cunha SC, Fernandes JO, Jager A, Petta T, et al. Assessment of multiple mycotoxins in breakfast cereals available in the Portuguese market. Food Chemistry. 2018;239:132-40. 48. Rodrigues I, Naehrer K. A three-year survey on the worldwide occurrence of mycotoxins in feedstuffs and feed. Toxins. 2012;4(9):663-75. 49. Gareis M, Scheuer R. Prevention of mycotoxin contamination of meat and meat products. Proceedings of International Symposium of Mycotoxicology ’99 Mycotoxin Contamination: Health Risk and Prevention Project 1999:101–7. 50. Loi M, Fanelli F, Liuzzi VC, Logrieco AF, Mulè G. Mycotoxin Biotransformation by Native and Commercial Enzymes: Present and Future Perspectives. Toxins. 2017;9(4):111. 51. Abd-Elghany SM, Sallam KI. Rapid determination of total aflatoxins and ochratoxins A in meat products by immuno-affinity fluorimetry. Food Chem. 2015;179:253-6. 52. Cao X, Li X, Li J, Niu Y, Shi L, Fang Z, et al. Quantitative determination of carcinogenic mycotoxins in human and animal biological matrices and animal-derived foods using multi-mycotoxin and analyte-specific high performance liquid chromatography- tandem mass spectrometric methods. Journal of chromatography B, Analytical technologies in the biomedical and life sciences. 2018;1073:191-200. 53. Iqbal SZ, Nisar S, Asi MR, Jinap S. Natural incidence of aflatoxins, ochratoxin A and zearalenone in chicken meat and eggs. Food Control. 2014;43:98-103. 54. Zhao Z, Liu N, Yang L, Deng Y, Wang J, Song S, et al. Multi-mycotoxin analysis of animal feed and animal-derived food using LC-MS/MS system with timed and highly selective reaction monitoring. Analytical and bioanalytical chemistry. 2015;407(24):7359-68. 55. Zou Z, He Z, Li H, Han P, Tang J, Xi C, et al. Development and application of a method for the analysis of two trichothecenes: deoxynivalenol and T-2 toxin in meat in China by HPLC-MS/MS. Meat science. 2012;90(3):613-7. 56. Cavus S, Tornuk F, Sarioglu K, Yetim H. Determination of mold contamination and aflatoxin levels of the meat products/ingredients collected from Turkey market. Journal of Food Safety. 2018;38(5):e12494. 57. Custódio L, Prados LF, Oliveira IM, Resende FD, Pettigrew JE, Yiannikouris A, et al. Do mycotoxin contaminated diets and yeast cell wall adsorbent affect meat quality of Nellore bulls finished in feedlot? - A short communication. Meat science. 2019;158:107865.

174

REFERENCES

58. Wang RJ, Fui SX, Miao CH, Feng DY. Effects of different mycotoxin adsorbents on performance, meat characteristics and blood profiles of avian broilers fed mold contaminated corn. Asian-Australasian Journal of Animal Sciences. 2006;19(1):72-9. 59. Surai PF, Dvorska JE, Sparks N, Jacques K. Proceedings of Alltech's 18th international symposium. T.P. Lyons, K.A. Jacques (Eds.), Impact of mycotoxins on the body's antioxidant defense, Nottingham University Press, Nottingham (2002), pp. 131-141. 2002. 60. EC. European Commission Regulation (EC) No. 1126/2007 of 28 September 2007 amending Regulation (EC) No. 1881/2006 setting maximum levels for certain contaminants in foodstuffs as regards Fusarium toxins in maize and maize products. . Official Journal of European Union L 255/14. 2007. 61. EC. European Commission regulation (EU) no 165/2010 of 26 February 2010, amending regulation (EC) no 1881/2006 setting maximum levels for certain contaminants in foodstuffs as regards aflatoxin. . Official Journal of the European Union,8e12 L 50. 2010. 62. Refai M, Niazi Z, Aziz N, Khafaga N. Incidence of aflatoxin B1 in the Egyptain cured meat basterma and control by γ-irradiation. Food/Nahrung. 2003;47:377-82. 63. Mottram DS. The Maillard Reaction: Source of Flavour in Thermally Processed Foods. In: Flavours and Fragrances: Chemistry, Bioprocessing and Sustainability; Berger, R. G., Ed.; Springer Berlim, Germany. p. 269-2832007. 64. Perez-Palacios T, Petisca C, Casal S, Ferreira IM. Changes in chemical composition of frozen coated fish products during deep-frying. Int J Food Sci Nutr. 2014;65(2):212- 8. 65. Vadivambal R, Jayas DS. Non-uniform Temperature Distribution During Microwave Heating of Food Materials—A Review. Food and Bioprocess Technology. 2010;3(2):161-71. 66. Ferreira VC, Morcuende D, Madruga MS, Hernandez-Lopez SH, Silva FA, Ventanas S, et al. Effect of pre-cooking methods on the chemical and sensory deterioration of ready-to-eat chicken patties during chilled storage and microwave reheating. Journal of food science and technology. 2016;53(6):2760-9. 67. Unlusayin M, Kaleli S, Bilgin S, Izci L, Gunlu A. Effect of hot-smoked rainbow trout (Oncorhynchus mykiss) diets on the protein levels of liver tissue in rats. Ann Nutr Metab. 2006;50(1):1-6. 68. Baldwin DE. Sous vide cooking: A review. International Journal of Gastronomy and Food Science. 2012;1(1):15-30. 69. Carmody RN, Weintraub GS, Wrangham W. Energetic consequences of thermal and nonthermal food processing. Proc Natl Acad Sci 2011;108(48):19199-203. 70. Van Boekel M, Fogliano V, Pellegrini N, Stanton C, Scholz G, Lalljie S, et al. A review on the beneficial aspects of food processing. Mol Nutr Food Res. 2010;54:1215-47. 71. Davey CL, Gilbert KV. Temperature-dependent cooking toughness in beef. J Sci Food Agric. 1974;25(8):931-8. 72. Meade SJ, Reid EA, Gerrard JA. The impact of processing on the nutritional quality of food proteins. Journal of AOAC International. 2005;88(3):904-22. 73. Tornberg E. Effect of heat on meat proteins- Implications on structure and quality of meat products: A review. Meat science. 2005;70:493-508. 74. Yu T-Y, Morton JD, Clerens S, Dyer JM. Cooking-Induced Protein Modifications in Meat. Comprehensive Reviews in Food Science and Food Safety. 2017;16(1):141- 59. 75. Alfaia CM, Alves SP, Lopes AF, Fernandes MJ, Costa AS, Fontes CM, et al. Effect of cooking methods on fatty acids, conjugated isomers of linoleic acid and nutritional quality of beef intramuscular fat. Meat science. 2010;84(4):769-77. 76. Domínguez R, Borrajo P, Lorenzo JM. The effect of cooking methods on nutritional value of foal meat. Journal of Food Composition and Analysis. 2015;43:61-7.

175

REFERENCES

77. Gerber N, Scheeder MR, Wenk C. The influence of cooking and fat trimming on the actual nutrient intake from meat. Meat science. 2009;81(1):148-54. 78. Lopes AF, Alfaia CMM, Partidário AMCPC, Lemos JPC, Prates JAM. Influence of household cooking methods on amino acids and minerals of Barrosã-PDO veal. Meat science. 2015;99:38-43. 79. Wilkinson BH, Lee E, Purchas RW, Morel PC. The retention and recovery of amino acids from pork longissimus muscle following cooking to either 60 degrees C or 75 degrees C. Meat science. 2014;96(1):361-5. 80. Ersoy B, Özeren A. The effect of cooking methods on mineral and vitamin contents of African catfish. Food Chem. 2009;115(2):419-22. 81. Hosseini H, Mahmoudzadeh M, Rezaei M, Mahmoudzadeh L, Khaksar R, Khosroshahi NK, et al. Effect of different cooking methods on minerals, vitamins and nutritional quality indices of kutum roach (Rutilus frisii kutum). Food Chem. 2014;148:86-91. 82. Jensen IJ, Dort J, Eilertsen KE. Proximate composition, antihypertensive and antioxidative properties of the semimembranosus muscle from pork and beef after cooking and in vitro digestion. Meat science. 2014;96(2 Pt A):916-21. 83. Karimian-Khosroshahi N, Hosseini H, Rezaei M, Khaksar R, Mahmoudzadeh M. Effect of Different Cooking Methods on Minerals, Vitamins, and Nutritional Quality Indices of Rainbow Trout (Oncorhynchus mykiss). Int J Food Prop. 2015;19(11):2471- 80. 84. Ortigues-Marty I, Thomas E, Preveraud DP, Girard CL, Bauchart D, Durand D, et al. Influence of maturation and cooking treatments on the nutritional value of bovine meats: Water losses and vitamin B12. Meat science. 2006;73(3):451-8. 85. Soladoye OP, Juárez ML, Aalhus JL, Shand P, Estévez M. Protein Oxidation in Processed Meat: Mechanisms and Potential Implications on Human Health. Comprehensive Reviews in Food Science and Food Safety. 2015;14(2):106-22. 86. Sante-Lhoutellier V, Astruc T, Marinova P, Greve E, Gatellier P. Effect of meat cooking on physicochemical state and in vitro digestibility of myofibrillar proteins. Journal of agricultural and food chemistry. 2008;56(4):1488-94. 87. Traore S, Aubry L, Gatellier P, Przybylski W, Jaworska D, Kajak-Siemaszko K, et al. Effect of heat treatment on protein oxidation in pig meat. Meat science. 2012;91(1):14- 21. 88. Cui C, Zhou X, Zhao M, Yang B. Effect of thermal treatment on the enzymatic hydrolysis of chicken proteins. Innovative Food Science & Emerging Technologies. 2009;10(1):37-41. 89. Gatellier P, Kondjoyan A, Portanguen S, Sante-Lhoutellier V. Effect of cooking on protein oxidation in n-3 polyunsaturated fatty acids enriched beef. Implication on nutritional quality. Meat science. 2010;85(4):645-50. 90. Roldan M, Antequera T, Armenteros M, Ruiz J. Effect of different temperature-time combinations on lipid and protein oxidation of sous-vide cooked lamb loins. Food Chem. 2014;149:129-36. 91. Davies MJ. Protein oxidation and peroxidation. The Biochemical journal. 2016;473(7):805-25. 92. Utrera M, Estevez M. Oxidative damage to poultry, pork, and beef during frozen storage through the analysis of novel protein oxidation markers. Journal of agricultural and food chemistry. 2013;61(33):7987-93. 93. Li, Xu Y, Li J, Niu S, Wang C, Zhang N, et al. Effect of oxidized lipids stored under different temperatures on muscle protein oxidation in Sichuan-style sausages during ripening. Meat science. 2019;147:144-54. 94. Gatellier P, Santé-Lhoutellier V, Portanguen S, Kondjoyan A. Use of meat fluorescence emission as a marker of oxidation promoted by cooking. Meat science. 2009;83(4):651-6.

176

REFERENCES

95. Hernandez-Lopez SH, Rodriguez-Carpena JG, Lemus-Flores C, Galindo-Garcia J, Estevez M. Antioxidant protection of proteins and lipids in processed pork loin chops through feed supplementation with avocado. Journal of food science and technology. 2016;53(6):2788-96. 96. Domínguez R, Gómez M, Fonseca S, Lorenzo JM. Influence of thermal treatment on formation of volatile compounds, cooking loss and lipid oxidation in foal meat. LWT - Food Science and Technology. 2014;58(2):439-45. 97. Pérez-Palacios T, Casal S, Petisca C, Ferreira IMPLVO. Nutritional and Sensory Characteristics of Bread-Coated Hake Fillets as Affected by Cooking Conditions. Journal of Food Quality. 2013;36(6):375-84. 98. Estevez M. Oxidative damage to poultry: from farm to fork. Poultry science. 2015;94(6):1368-78. 99. Min B, Nam KC, Cordray J, Ahn DU. Endogenous factors affecting oxidative stability of beef loin, pork loin, and chicken breast and thigh meats. J Food Sci. 2008;73(6):C439-46. 100. Nieva-Echevarria B, Goicoechea E, Guillen MD. Food lipid oxidation under gastrointestinal digestion conditions: A review. Critical reviews in food science and nutrition. 2018:1-18. 101. Ayala A, Munoz MF, Arguelles S. Lipid peroxidation: production, metabolism, and signaling mechanisms of malondialdehyde and 4-hydroxy-2-nonenal. Oxidative medicine and cellular longevity. 2014;2014:360438. 102. Hu, Ren S, Shen Q, Chen J, Ye X, Ling J. Proteomic study of the effect of different cooking methods on protein oxidation in fish fillets. RSC Advances. 2017;7(44):27496-505. 103. Pérez Palacios T. Effect of Cooking Conditions on Quality Characteristics of Confit Cod: Prediction by MRI. International journal of food engineering. 2017;v. 13(no. 8):pp. 15-2017 v.13 no.8. 104. Lešková E, Kubíková J, Kováčiková E, Košická M, Porubská J, Holčíková K. Vitamin losses: Retention during heat treatment and continual changes expressed by mathematical models. J Food Compos Anal. 2006;19(4):252-76. 105. Lund D. Effects of Heat Processing on Nutrients. In: Karmas E, Harris RS, editors. Nutritional Evaluation of Food Processing. Dordrecht: Springer Netherlands; 1988. p. 319-54. 106. Kumar S, Aalbersberg B. Nutrient retention in foods after earth-oven cooking compared to other forms of domestic cooking. J Food Compos Anal. 2006;19(4):311- 20. 107. Czerwonka M, Szterk A, Waszkiewicz-Robak B. Vitamin B12 content in raw and cooked beef. Meat science. 2014;96(3):1371-5. 108. Bolger Z, Brunton NP, Lyng JG, Monahan FJ. Quality attributes and retention of vitamin E in reduced salt chicken sausages fortified with vitamin E. Journal of food science and technology. 2016;53(11):3948-59. 109. Tomović VM, Vujadinović DD, Grujić RP, Jokanović MR, Kevrešan ŽS, Škaljac SB, et al. Effect of Endpoint Internal Temperature on Mineral Contents of Boiled Pork Loin. Journal of Food Processing and Preservation. 2015;39(6):1854-8. 110. Oz F, Aksu MI, Turan M. The Effects of Different Cooking Methods on Some Quality Criteria and Mineral Composition of Beef Steaks. Journal of Food Processing and Preservation. 2017;41(4):e13008. 111. Purchas RW, Wilkinson BHP, Carruthers F, Jackson F. A comparison of the nutrient content of uncooked and cooked lean from New Zealand beef and lamb. Journal of Food Composition and Analysis. 2014;35(2):75-82. 112. Costa M, Viegas O, Melo A, Petisca C, Pinho O, Ferreira IMPLVO. Heterocyclic Aromatic Amine Formation in Barbecued Sardines (Sardina pilchardus) and Atlantic Salmon (Salmo salar). J Agric Food Chem. 2009;57:3173-9.

177

REFERENCES

113. Melo A, Viegas O, Eça R, Petisca C, Pinho O, Ferreira IMPLVO. Extraction, Detection, and Quantification of Heterocyclic Aromatic Amines in Portuguese Meat Dishes by HPLC/Diode Array. Journal of Liquid Chromatography & Related Technologies. 2008;31(5):772-87. 114. Rahman Uu, Sahar A, Khan MI, Nadeem M. Production of heterocyclic aromatic amines in meat: Chemistry, health risks and inhibition. A review. LWT - Food Science and Technology. 2014;59(1):229-33. 115. Viegas, Amaro LF, Ferreira IMPLVO, Pinho O. Inhibitory Effect of Antioxidant-Rich Marinades on the Formation of Heterocyclic Aromatic Amines in Pan-Fried Beef. Journal of Agricultural and Food Chemistry. 2012;60(24):6235-40. 116. Garcia-Lomillo J, Viegas O, Gonzalez-SanJose ML, Ferreira IM. Influence of red wine pomace seasoning and high-oxygen atmosphere storage on carcinogens formation in barbecued beef patties. Meat science. 2017;125:10-5. 117. Viegas O, Novo P, Pinho O, Ferreira IM. A comparison of the extraction procedures and quantification methods for the chromatographic determination of polycyclic aromatic hydrocarbons in charcoal grilled meat and fish. Talanta. 2012;88:677-83. 118. John EM, Stern MC, Sinha R, Koo J. Meat consumption, cooking practices, meat mutagens, and risk of prostate cancer. Nutr Cancer. 2011;63(4):525-37. 119. Zhang Y, Yu C, Mei J, Wang S. Formation and mitigation of heterocyclic aromatic amines in fried pork. Food additives & contaminants Part A, Chemistry, analysis, control, exposure & risk assessment. 2013;30(9):1501-7. 120. Shabbir MA, Raza A, Anjum FM, Khan MR, Suleria HA. Effect of thermal treatment on meat proteins with special reference to heterocyclic aromatic amines (HAAs). Critical reviews in food science and nutrition. 2015;55(1):82-93. 121. Ni W, McNaughton L, LeMaster DM, Sinha R, Turesky RJ. Quantitation of 13 Heterocyclic Aromatic Amines in Cooked Beef, Pork, and Chicken by Liquid Chromatography-Electrospray Ionization/Tandem Mass Spectrometry. Journal of agricultural and food chemistry. 2008;56:68-78. 122. Polak T, Andrenšek S, Žlender B, Gašperlin L. Effects of ageing and low internal temperature of grilling on the formation of heterocyclic amines in beef Longissimus dorsi muscle. LWT - Food Science and Technology. 2009;42(1):256-64. 123. Skog KI, Johansson MA, Jagerstad MI. Carcinogenic heterocyclic amines in model systems and cooked foods: a review on formation, occurrence and intake. Food and chemical toxicology : an international journal published for the British Industrial Biological Research Association. 1998;36(9-10):879-96. 124. Oz F, Kızıl M, Çelık T. Effects of Different Cooking Methods on the Formation of Heterocyclic Aromatic Amines in Goose Meat. Journal of Food Processing and Preservation. 2016;40(5):1047-53. 125. Wang Y, Hui T, Zhang YW, Liu B, Wang FL, Li JK, et al. Effects of frying conditions on the formation of heterocyclic amines and trans fatty acids in grass carp (Ctenopharyngodon idellus). Food Chem. 2015;167:251-7. 126. Farhadian A, Jinap S, Hanifah HN, Zaidul IS. Effects of meat preheating and wrapping on the levels of polycyclic aromatic hydrocarbons in charcoal-grilled meat. Food Chem. 2011;124(1):141-6. 127. Farhadian A, Jinap S, Abas F, Sakar ZI. Determination of polycyclic aromatic hydrocarbons in grilled meat. Food Control. 2010;21(5):606-10. 128. Aaslyng MD, Duedahl-Olesen L, Jensen K, Meinert L. Content of heterocyclic amines and polycyclic aromatic hydrocarbons in pork, beef and chicken barbecued at home by Danish consumers. Meat science. 2013;93(1):85-91. 129. Xia Z, Duan X, Qiu W, Liu D, Wang B, Tao S, et al. Health risk assessment on dietary exposure to polycyclic aromatic hydrocarbons (PAHs) in Taiyuan, China. Sci Total Environ. 2010;408(22):5331-7.

178

REFERENCES

130. JECFA. Joint fao/who expert committee on food additives. 2005. Sixty-fourth meeting. Rome, 8–17 february 2005. Summary and conclusions. Available from: Http://www.Fao.Org/3/a-at877e.Pdf (accessed 2017 april 28th). 2005. 131. EFSA. Scientific opinion of the panel on contaminants in the food chain on a request from the european commission on polycyclic aromatic hydrocarbons in food. EFSA J 724, 1–114. 2008. 132. Bansal V, Kim KH. Review of PAH contamination in food products and their health hazards. Environ Int. 2015;84:26-38. 133. Kim KH, Jahan SA, Kabir E, Brown RJ. A review of airborne polycyclic aromatic hydrocarbons (PAHs) and their human health effects. Environ Int. 2013;60:71-80. 134. Ravindra K, Sokhi R, Vangrieken R. Atmospheric polycyclic aromatic hydrocarbons: Source attribution, emission factors and regulation. Atmospheric Environment. 2008;42(13):2895-921. 135. Ledesma E, Rendueles M, Díaz M. Contamination of meat products during smoking by polycyclic aromatic hydrocarbons: Processes and prevention. Food Control. 2016;60:64-87. 136. Gorji MEh, Ahmadkhaniha R, Moazzen M, Yunesian M, Azari A, Rastkari N. Polycyclic aromatic hydrocarbons in Iranian Kebabs. Food Control. 2016;60:57-63. 137. Chung SY, Yettella RR, Kim JS, Kwon K, Kim MC, Min DB. Effects of grilling and roasting on the levels of polycyclic aromatic hydrocarbons in beef and pork. Food Chem. 2011;129(4):1420-6. 138. Heshmati A. Impact of Cooking Procedures on Antibacterial Drug Residues in Foods: A Review. J Food Qual Hazards Control. 2015;2:33-7. 139. Tian L, Khalil S, Bayen S. Effect of thermal treatments on the degradation of antibiotic residues in food. Critical reviews in food science and nutrition. 2017;57(17):3760-70. 140. Abou-raya SH, Shalaby AR, Salama NA, Emam WH, Mehaya FM. Effect of Ordinary Cooking Procedures on Tetracycline Residues in Chicken Meat. J Food Drug Anal. 2013;21(1):80-6. 141. Javadi A. Effect of roasting, boiling and microwaving cooking method on Doxycline residues in edible tissues of poultry by microbial method. Afr J Pharm Pharmacol. 2011;5(8):1034-7. 142. Hussein MA, Ahmed MM, Morshedy AM. Effect of cooking methods on some antibiotic residues in chicken meat. Jpn J Vet Res 2016;64 (Supplement 2):S225-S31. 143. Loftin KA, Adams CD, Meyer MT, Surampalli R. Effects of ionic strength, temperature, and pH on degradation of selected antibiotics. Journal of environmental quality. 2008;37(2):378-86. 144. Gratacos-Cubarsi M, Fernandez-Garcia A, Picouet P, Valero-Pamplona A, Garcia- Regueiro JA, Castellari M. Formation of tetracycline degradation products in chicken and pig meat under different thermal processing conditions. Journal of agricultural and food chemistry. 2007;55(11):4610-6. 145. Moreno-Cerezo JM, Cordoba-Diaz M, Cordoba-Diaz D, Cordoba-Borrego M. A stability study of tetracycline and tetracycline cyclodextrins in tablets using a new HPLC method. Journal of pharmaceutical and biomedical analysis. 2001;26(3):417- 26. 146. Nguyen V, Li C, Zhou G. The degradation of oxytetracycline during thermal treatments of chicken and pig meat and the toxic effects of degradation products of oxytetracycline on rats. Journal of food science and technology. 2015;52(5):2842-50. 147. Hsieh MK, Shyu CL, Liao JW, Franje CA, Huang YJ, Chang SK, et al. Correlation analysis of heat stability of veterinary antibiotics by structural degradation, changes in antimicrobial activity and genotoxicity. Vet Med. 2011;56(6):274-85. 148. Javadi A, Mirzaei H, Khatibi SA. Effect of Roasting Process on Antibiotic Residues in Edible Tissues of Poultry by FPT method. J Anim Vet Adv. 2009;8(12):2468-72.

179

REFERENCES

149. Sattar S, Hassan MM, M ISK, Alam M, Faruk MSA, Chowdhury S, et al. Antibiotic residues in broiler and layer meat in chittagong district of Bangladesh. Vet World 2014;7(9):738-43. 150. Salaramoli J, Heshmati A, Kamkar A, Hassan J. Effect of cooking procedures on tylosin residues in chicken meatball. Journal für Verbraucherschutz und Lebensmittelsicherheit. 2016;11(1):53-60. 151. Fathy F, Ahmed AM, Moursi MK. Effect of Cooking Methods on Antibiotic Residues in Broiler Chicken Meat. 2nd Conference of Food Safety. 2015;I:76-81. 152. Lolo M, Pedreira S, Miranda JM, Vazquez BI, Franco CM, Cepeda A, et al. Effect of cooking on enrofloxacin residues in chicken tissue. Food additives and contaminants. 2006;23(10):988-93. 153. Ismail-Fitry MR, Jinap S, Jamilah B, Saleha AA. Effect of deep-frying at different temperature and time on sulfonamide residues in chicken meat-balls. J Food Drug Anal. 2008;16:81-6. 154. Ismail-Fitry MR, Jinap S, Jamilah B, Saleha AA. Effect of Different Time and Temperature of Various Cooking Methods on Sulfonamide Residues in Chicken Balls. In: Hüseyin Gökçekus, Umut Türker, James W. LaMoreaux. Survival and Sustainability. 2010:607-13. 155. Cooper KM, Kennedy DG. Stability studies of the metabolites of nitrofuran antibiotics during storage and cooking. Food additives and contaminants. 2007;24(9):935-42. 156. Javadi A, Mirzaie H, Khatibi SA. Effect of roasting, boiling and microwaving cooking method on sulfadiazine + trimethoprim residues in edible tissues of broiler by microbial inhibition method. Afr J Microbiol Res. 2011;5:96-9. 157. Steffenak I, Hormazabal V, Yndestad M. Reservoir of quinolone residues in fish. Food Additives & Contaminants. 1991;8(6):777-80. 158. Shitandi AA, Aila O, Ottaro S, Aliong’o L, Mwangi G, Kumar- Sharma H, et al. Effect of deep frying on furazolidone anticoccidial drug residues in liver and muscle tissues of chicken African Journal of Food Science. 2008;2:144-8. 159. Karlovsky P, Suman M, Berthiller F, De Meester J, Eisenbrand G, Perrin I, et al. Impact of food processing and detoxification treatments on mycotoxin contamination. Mycotoxin Research. 2016;32(4):179-205. 160. Milani J, Maleki G. Effects of processing on mycotoxin stability in cereals. Journal of the science of food and agriculture. 2014;94(12):2372-5. 161. Aiko V, Mehta A. Occurrence, detection and detoxification of mycotoxins. Journal of biosciences. 2015;40(5):943-54. 162. Kabak. The fate of mycotoxins during thermal food processing. Journal of the Science of Food and Agriculture. 2009;89(4):549-54. 163. Schmidt HS, Becker S, Cramer B, Humpf H-U. Impact of Mechanical and Thermal Energies on the Degradation of T-2 and HT-2 Toxins during Extrusion Cooking of Oat Flour Henning Sören Schmidt, Stefanie Becker, Benedikt Cramer, and Hans-Ulrich Humpf. Journal of Agricultural and Food Chemistry. 2017;65(20):4177-83. 164. Furtado RM, Pearson AM, Gray JI, Hogberg MG, Miller ER. Effects of Cooking and/or Processing Upon Levels of Aflatoxins in Meat from Pigs Fed A Contaminated Diet. Journal of Food Science. 1981;46(5):1306-8. 165. Tolosa J, Font G, Manes J, Ferrer E. Mitigation of enniatins in edible fish tissues by thermal processes and identification of degradation products. Food and chemical toxicology : an international journal published for the British Industrial Biological Research Association. 2017;101:67-74. 166. Mackie A. The digestive tract: A complex system. Interdisciplinary Approaches to Food Digestion2019. p. 11-27. 167. Cardoso C, Afonso C, Lourenço H, Costa S, Nunes ML. Bioaccessibility assessment methodologies and their consequences for the risk–benefit evaluation of food. Trends in Food Science & Technology. 2015;41(1):5-23. 180

REFERENCES

168. Afonso C, Cardoso C, Freire M, Silva IE, Linares F, Villanueva JLR, et al. The impact of alternative dietary lipids on the in vitro bioaccessibility of sole fillets for human consumption. Aquaculture. 2017;474:66-74. 169. Lucas-González R, Viuda-Martos M, Pérez-Alvarez JA, Fernández-López J. In vitro digestion models suitable for foods: Opportunities for new fields of application and challenges. Food Research International. 2018;107:423-36. 170. Steppeler C, Haugen JE, Rodbotten R, Kirkhus B. Formation of Malondialdehyde, 4- Hydroxynonenal, and 4-Hydroxyhexenal during in Vitro Digestion of Cooked Beef, Pork, Chicken, and Salmon. Journal of agricultural and food chemistry. 2016;64(2):487-96. 171. Tullberg C, Larsson K, Carlsson NG, Comi I, Scheers N, Vegarud G, et al. Formation of reactive aldehydes (MDA, HHE, HNE) during the digestion of cod liver oil: comparison of human and porcine in vitro digestion models. Food & function. 2016;7(3):1401-12. 172. Van Hecke T, Ho PL, Goethals S, De Smet S. The potential of herbs and spices to reduce lipid oxidation during heating and gastrointestinal digestion of a beef product. Food research international (Ottawa, Ont). 2017;102:785-92. 173. Gobert M, Rémond D, Loonis M, Buffière C, Santé-Lhoutellier V, Dufour C. Fruits, vegetables and their polyphenols protect dietary lipids from oxidation during gastric digestion. Food & function. 2014;5(9):2166-74. 174. Tirosh O, Shpaizer A, Kanner J. Lipid Peroxidation in a Stomach Medium Is Affected by Dietary Oils (Olive/Fish) and Antioxidants: The Mediterranean versus Western Diet. Journal of agricultural and food chemistry. 2015;63(31):7016-23. 175. Gorelik S, Ligumsky M, Kohen R, Kanner J. The Stomach as a “Bioreactor”: When Red Meat Meets Red Wine. Journal of Agricultural and Food Chemistry. 2008;56(13):5002-7. 176. Kanner J, Lapidot T. The stomach as a bioreactor: dietary lipid peroxidation in the gastric fluid and the effects of plant-derived antioxidants. Free Radical Biology and Medicine. 2001;31(11):1388-95. 177. Gueraud F, Tache S, Steghens JP, Milkovic L, Borovic-Sunjic S, Zarkovic N, et al. Dietary polyunsaturated fatty acids and heme iron induce oxidative stress biomarkers and a cancer promoting environment in the colon of rats. Free radical biology & medicine. 2015;83:192-200. 178. Sirota R, Gorelik S, Harris R, Kohen R, Kanner J. Coffee polyphenols protect human plasma from postprandial carbonyl modifications. Molecular nutrition & food research. 2013;57(5):916-9. 179. Bax ML, Aubry L, Ferreira C, Daudin JD, Gatellier P, Remond D, et al. Cooking temperature is a key determinant of in vitro meat protein digestion rate: investigation of underlying mechanisms. Journal of agricultural and food chemistry. 2012;60(10):2569-76. 180. Sun W, Zhou F, Zhao M, Yang B, Cui C. Physicochemical changes of myofibrillar proteins during processing of Cantonese sausage in relation to their aggregation behaviour and in vitro digestibility. Food Chemistry. 2011;129(2):472-8. 181. Versantvoort CH, van de Kamp E, Rompelberg CJ. Development and applicability of an in vitro digestion model in assessing the bioaccessibility of contaminants from food. RIVM report 320102002. 2004. 182. Glanzer S, Pulido SA, Tutz S, Wagner GE, Kriechbaum M, Gubensak N, et al. Structural and functional implications of the interaction between macrolide antibiotics and bile acids. Chemistry (Weinheim an der Bergstrasse, Germany). 2015;21(11):4350-8. 183. Hörter D, Dressman JB. Influence of physicochemical properties on dissolution of drugs in the gastrointestinal tract1PII of original article: S0169-409X(96)00487-5. The article was originally published in Advanced Drug Delivery Reviews 25 (1997) 3–14.1. Advanced Drug Delivery Reviews. 2001;46(1):75-87.

181

REFERENCES

184. Welling PG. Influence of food and diet on gastrointestinal drug absorption: A review. Journal of Pharmacokinetics and Biopharmaceutics 1977;5:291. https://doi.org/10.1007/BF01061694. 185. Assuncao R, Martins C, Dupont D, Alvito P. Patulin and ochratoxin A co-occurrence and their bioaccessibility in processed cereal-based foods: A contribution for Portuguese children risk assessment. Food and chemical toxicology : an international journal published for the British Industrial Biological Research Association. 2016;96:205-14. 186. De Angelis E, Monaci L, Mackie A, Salt L, Visconti A. Bioaccessibility of T-2 and HT- 2 toxins in mycotoxin contaminated bread models submitted to in vitro human digestion. Innovative Food Science & Emerging Technologies. 2014;22:248-56. 187. González-Arias CA, Marín S, Sanchis V, Ramos AJ. Mycotoxin bioaccessibility/absorption assessment using in vitro digestion models: a review. World Mycotoxin Journal. 2013;6(2):167-84. 188. Kabak B, Ozbey F. Assessment of the bioaccessibility of aflatoxins from various food matrices using an in vitro digestion model, and the efficacy of probiotic bacteria in reducing bioaccessibility. Journal of Food Composition and Analysis. 2012;27(1):21- 31. 189. Versantvoort CH, Oomen AG, Van de Kamp E, Rompelberg CJ, Sips AJ. Applicability of an in vitro digestion model in assessing the bioaccessibility of mycotoxins from food. Food and chemical toxicology : an international journal published for the British Industrial Biological Research Association. 2005;43(1):31-40. 190. Ferrer M, Manyes L, Manes J, Meca G. Influence of prebiotics, probiotics and protein ingredients on mycotoxin bioaccessibility. Food & function. 2015;6(3):987-94. 191. Brodkorb A, Egger L, Alminger M, Alvito P, Assunção R, Ballance S, et al. INFOGEST static in vitro simulation of gastrointestinal food digestion. Nature Protocols. 2019;14(4):991-1014. 192. Minekus M, Alminger M, Alvito P, Ballance S, Bohn T, Bourlieu C, et al. A standardised static in vitro digestion method suitable for food - an international consensus. Food & function. 2014;5(6):1113-24. 193. Dupont D, Alric M, Blanquet-Diot S, Bornhorst G, Cueva C, Deglaire A, et al. Can dynamic in vitro digestion systems mimic the physiological reality? Critical reviews in food science and nutrition. 2019;59(10):1546-62. 194. Rodrigues DB, Mariutti LRB, Mercadante AZ. An: In vitro digestion method adapted for carotenoids and carotenoid esters: Moving forward towards standardization. Food and Function. 2016;7(12):4992-5001. 195. Fernández-García E, Carvajal-Lérida I, Pérez-Gálvez A. In vitro bioaccessibility assessment as a prediction tool of nutritional efficiency. Nutrition Research. 2009;29(11):751-60. 196. Yang Y, Yu S, Tan Y, Liu N, Wu A. Individual and Combined Cytotoxic Effects of Co- Occurring Deoxynivalenol Family Mycotoxins on Human Gastric Epithelial Cells. Toxins. 2017;9(3). 197. Sarries MV, Murray BE, Moloney AP, Troy D, Beriain MJ. The effect of cooking on the fatty acid composition of longissimus muscle from beef heifers fed rations designed to increase the concentration of conjugated linoleic acid in tissue. Meat science. 2009;81(2):307-12. 198. Santos T, Lourenço BN, Coentro J, Granja PL. Cell-based in vitro models for gastric permeability studies. In Sarmento B (ed) Concepts and Modela for Drug Permeability studies, 1st edition. Elsevier, pp 41-56. 2016. 199. Pereira C, Costa J, Sarmento B, Araújo F. Cell-based in vitro models for intestinal permeability studies. In Sarmento B (ed) Concepts and Models for Drug Permeability studies, 1st edition. Elsevier, pp 57-81. 2016. 200. Hessel S, Lampen A, Seidel A. Polycyclic aromatic hydrocarbons in food--efflux of the conjugated biomarker 1-hydroxypyrene is mediated by Breast Cancer Resistance

182

REFERENCES

Protein (ABCG2) in human intestinal Caco-2 cells. Food and chemical toxicology : an international journal published for the British Industrial Biological Research Association. 2013;62:797-804. 201. Cabrera-Pérez MA, Sanz MB, Sanjuan VM, González-Álvarez M, Álvarez IG. Importance and applications of cell- and tissue-based in vitro models for drug permeability screening in early stages of drug development. In Sarmento B (ed) Concepts and Modela for Drug Permeability studies, 1st edition. Elsevier, pp 3-29. 2016. 202. Maresca M, Fantini J. Some food-associated mycotoxins as potential risk factors in humans predisposed to chronic intestinal inflammatory diseases. Toxicon : official journal of the International Society on Toxinology. 2010;56(3):282-94. 203. Awada M, Soulage CO, Meynier A, Debard C, Plaisancié P, Benoit B, et al. Dietary oxidized n-3 PUFA induce oxidative stress and inflammation: role of intestinal absorption of 4-HHE and reactivity in intestinal cells. Journal of lipid research. 2012;53(10):2069-80. 204. Maestre R, Douglass JD, Kodukula S, Medina I, Storch J. Alterations in the Intestinal Assimilation of Oxidized PUFAs Are Ameliorated by a Polyphenol-Rich Grape Seed Extract in an In Vitro Model and Caco-2 Cells. The Journal of Nutrition. 2013;143(3):295-301. 205. Penumetcha M, Khan N, Parthasarathy S. Dietary oxidized fatty acids: an atherogenic risk? Journal of lipid research. 2000;41(9):1473-80. 206. Dasilva G, Boller M, Medina I, Storch J. Relative levels of dietary EPA and DHA impact gastric oxidation and essential fatty acid uptake. The Journal of Nutritional Biochemistry. 2018;55:68-75. 207. Qiu J, Zhao T, Liu Q, He J, He D, Wu G, et al. Residual veterinary antibiotics in pig excreta after oral administration of sulfonamides. Environmental Geochemistry and Health. 2016;38(2):549-56. 208. Agwuh KN, MacGowan A. Pharmacokinetics and pharmacodynamics of the tetracyclines including glycylcyclines. The Journal of antimicrobial chemotherapy. 2006;58(2):256-65. 209. Neuvonen PJ. Interactions with the absorption of tetracyclines. Drugs. 1976;11(1):45- 54. 210. Pitman SK, Hoang UTP, Wi CH, Alsheikh M, Hiner DA, Percival KM. Revisiting Oral Fluoroquinolone and Multivalent Cation Drug-Drug Interactions: Are They Still Relevant? . Antibiotics 2019;8:108. 211. Kivistö KT, Ojala-Karlsson P, Neuvonen PJ. Inhibition of norfloxacin absorption by dairy products. Antimicrob Agents Chemother. 1992;36(2):489-91. 212. Macdonald H, Place VA, Falk H, Darken MA. Effect of food on absorption of sulfonamides in man. Chemotherapy. 1967;12(5):282-5. 213. Grenier B, Applegate TJ. Modulation of intestinal functions following mycotoxin ingestion: meta-analysis of published experiments in animals. Toxins. 2013;5(2):396- 430. 214. Galtier P. Contribution of pharmacokinetic studies to mycotoxicology—Ochratoxin A. Veterinary Science Communications. 1977;1(1):349-58. 215. Rychlik M, Kircher F, Schusdziarra V, Lippl F. Absorption of the mycotoxin patulin from the rat stomach. Food and chemical toxicology : an international journal published for the British Industrial Biological Research Association. 2004;42(5):729- 35. 216. Hill NS, Thompson FN, Stuedemann JA, Rottinghaus GW, Ju HJ, Dawe DL, et al. Ergot alkaloid transport across ruminant gastric tissues. Journal of animal science. 2001;79(2):542-9. 217. Chain EPanel oCitF, Knutsen HK, Alexander J, Barregård L, Bignami M, Brüschweiler B, et al. Risks to human and animal health related to the presence of deoxynivalenol

183

REFERENCES

and its acetylated and modified forms in food and feed. EFSA Journal. 2017;15(9):e04718. 218. EFSA. Opinion of the scientific panel on contaminants in the food chain on a request from the commission related to ochratoxin A in food. The EFSA Journal. 2006;356:1- 56. 219. EFSA. Opinion of the scientific panel on contaminants in the food chain [CONTAM] related to the potential increase of consumer health risk by a possible increase of the existing maximum levels for aflatoxins in almonds, hazelnuts and pistachios and derived products. The EFSA Journal. 2007;446:1-127. 220. EC. European Commission. Opinion of the scientific committee on food on fusarium toxins part 3: Fumonisin B1 (FB1). 2000. 221. IARC. Monographs on the evaluation of carcinogenic risks to humans: some naturally occurring substances: food items and constituents, heterocyclic aromatic amines and mycotoxins. Lyon, France: IARC press; 1993. 222. IARC. Traditional herbal medicines, some mycotoxins, naphthalene and styrene. IARC Monographs on the evaluation of carcinogenic risks to humans. Lyon, France: IARC press; 2002. 223. IARC. Aflatoxins. Iarc Monogr Eval Carcinog Risks Humans. 2012; 100 F:225–48. 224. Knockaert G, Lemmens L, Van Buggenhout S, Hendrickx M, Van Loey A. Changes in β-carotene bioaccessibility and concentration during processing of carrot puree. Food Chemistry. 2012;133(1):60-7. 225. Sarmento B. Concepts and model for drug permeability studies. Cell and tissue based in vitro culture models. Cambridge, UK: Woodhead publishing series in biomedicine: number 79. Elsevier. . 2016. 226. Fernandes I, de Freitas V, Reis C, Mateus N. A new approach on the gastric absorption of anthocyanins. Food & function. 2012;3(5):508-16. 227. Oshima T, Miwa H, Joh T. Aspirin induces gastric epithelial barrier dysfunction by activating p38 MAPK via claudin-7. American journal of physiology Cell physiology. 2008;295(3):C800-6. 228. Basque JR, Chenard M, Chailler P, Menard D. Gastric cancer cell lines as models to study human digestive functions. Journal of cellular biochemistry. 2001;81(2):241-51. 229. Lemieux M, Bouchard F, Gosselin P, Paquin J, Mateescu MA. The NCI-N87 cell line as a gastric epithelial barrier model for drug permeability assay. Biochemical and biophysical research communications. 2011;412(3):429-34. 230. Akbari P, Braber S, Varasteh S, Alizadeh A, Garssen J, Fink-Gremmels J. The intestinal barrier as an emerging target in the toxicological assessment of mycotoxins. Archives of toxicology. 2017;91(3):1007-29. 231. Hubatsch I, Ragnarsson EGE, Artursson P. Determination of drug permeability and prediction of drug absorption in Caco-2 monolayers. Nature Protocols. 2007;2:2111. 232. Fogh J, Wright WC, Loveless JD. Absence of HeLa cell contamination in 169 cell lines derived from human tumors. Journal of the National Cancer Institute. 1977;58(2):209- 14. 233. Markowska M, Oberle R, Juzwin S, Hsu C-P, Gryszkiewicz M, Streeter AJ. Optimizing Caco-2 cell monolayers to increase throughput in drug intestinal absorption analysis. Journal of Pharmacological and Toxicological Methods. 2001;46(1):51-5. 234. Turco L, Catone T, Caloni F, Di Consiglio E, Testai E, Stammati A. Caco-2/TC7 cell line characterization for intestinal absorption: how reliable is this in vitro model for the prediction of the oral dose fraction absorbed in human? Toxicology in vitro : an international journal published in association with BIBRA. 2011;25(1):13-20. 235. Maubon N, Le Vee M, Fossati L, Audry M, Le Ferrec E, Bolze S, et al. Analysis of drug transporter expression in human intestinal Caco-2 cells by real-time PCR. Fundamental & clinical pharmacology. 2007;21(6):659-63. 236. Shimizu M. Interaction between food substances and the intestinal epithelium. Bioscience, biotechnology, and biochemistry. 2010;74(2):232-41.

184

REFERENCES

237. Sun H, Chow EC, Liu S, Du Y, Pang KS. The Caco-2 cell monolayer: usefulness and limitations. Expert opinion on drug metabolism & toxicology. 2008;4(4):395-411. 238. Sun H, Pang KS. Permeability, transport, and metabolism of solutes in Caco-2 cell monolayers: a theoretical study. Drug metabolism and disposition: the biological fate of chemicals. 2008;36(1):102-23. 239. Berger V, Gabriel A-F, Sergent T, Trouet A, Larondelle Y, Schneider Y-J. Interaction of ochratoxin A with human intestinal Caco-2 cells: possible implication of a multidrug resistance-associated protein (MRP2). Toxicology Letters. 2003;140-141:465-76. 240. Caloni F, Cortinovis C, Pizzo F, De Angelis I. Transport of Aflatoxin M(1) in Human Intestinal Caco-2/TC7 Cells. Frontiers in Pharmacology. 2012;3:111. 241. De Angelis I, Frigge G, Raimondi F, Stammati A, Zucco F, Caloni F. Absorption of fumonisin B1 and aminopentol on an in vitro model of intestinal epithelium; the role of P-glycoprotein. Toxicon : official journal of the International Society on Toxinology. 2005;45(3):285-91. 242. Kadota T, Furusawa H, Hirano S, Tajima O, Kamata Y, Sugita-Konishi Y. Comparative study of deoxynivalenol, 3-acetyldeoxynivalenol, and 15-acetyldeoxynivalenol on intestinal transport and IL-8 secretion in the human cell line Caco-2. Toxicology in vitro : an international journal published in association with BIBRA. 2013;27(6):1888- 95. 243. Sergent T, Parys M, Garsou S, Pussemier L, Schneider YJ, Larondelle Y. Deoxynivalenol transport across human intestinal Caco-2 cells and its effects on cellular metabolism at realistic intestinal concentrations. Toxicology letters. 2006;164(2):167-76. 244. Videmann B, Mazallon M, Tep J, Lecoeur S. Metabolism and transfer of the mycotoxin zearalenone in human intestinal Caco-2 cells. Food and chemical toxicology : an international journal published for the British Industrial Biological Research Association. 2008;46(10):3279-86. 245. Videmann B, Tep J, Cavret S, Lecoeur S. Epithelial transport of deoxynivalenol: Involvement of human P-glycoprotein (ABCB1) and multidrug resistance-associated protein 2 (ABCC2). Food and Chemical Toxicology. 2007;45(10):1938-47. 246. Mahler GJ, Shuler ML, Glahn RP. Characterization of Caco-2 and HT29-MTX cocultures in an in vitro digestion/cell culture model used to predict iron bioavailability. J Nutr Biochem. 2009;20(7):494-502. 247. Heys KA, Shore RF, Pereira MG, Jones KC, Martin FL. Risk assessment of environmental mixture effects. RSC Advances. 2016;6(53):47844-57. 248. EC. European commission (EC); commission regulation (EU). No 885/2010 of 7 october 2010 concerning the authorisation of the preparation of narasin and nicarbazin as a feed additive for chickens for fattening (holder of authorisation Eli Lilly and Company Ltd) and amending Regulation (EC) No 2430/1999. Off. J. Eur. Commun. 2010, 265, 5-8. 2010. 249. Bacila DM, Cunha A, Jr., Weber IF, Scheuermann GN, Coldebella A, Caron L, et al. Degradation of 4,4'-Dinitrocarbanilide in Chicken Breast by Thermal Processing. Journal of agricultural and food chemistry. 2018;66(31):8391-7. 250. IARC. Some Non-heterocyclic Polycyclic Aromatic Hydrocarbons and Some Related Exposures. IARC Monographs on the Evaluation of Carcinogenic Risks to Humans. Lyon, France: IARC press. 2010;92. 251. Guéraud F. 4-Hydroxynonenal metabolites and adducts in pre-carcinogenic conditions and cancer. Free Radical Biology and Medicine. 2017;111:196-208. 252. Łuczaj W, Gęgotek A, Skrzydlewska E. Antioxidants and HNE in redox homeostasis. Free Radical Biology and Medicine. 2017;111:87-101. 253. Csallany AS, Han I, Shoeman DW, Chen C, Yuan J. 4-Hydroxynonenal (HNE), a Toxic Aldehyde in French Fries from Fast Food Restaurants. Journal of the American Oil Chemists' Society. 2015;92(10):1413-9. 254. council BSH. Sécurité des huiles et des graisses, Avis n°8310. 2011.

185

REFERENCES

255. Kim J, Park SH, Do KH, Kim D, Moon Y. Interference with mutagenic aflatoxin B1- induced checkpoints through antagonistic action of ochratoxin A in intestinal cancer cells: a molecular explanation on potential risk of crosstalk between carcinogens. Oncotarget. 2016;7(26):39627-39. 256. Roy AK. Effects of aflatoxin B1 on polysomal profiles and RNA synthesis in rat liver. Biochimica et Biophysica Acta (BBA) - Nucleic Acids and Protein Synthesis. 1968;169(1):206-11. 257. Creppy EE, Chiarappa P, Baudrimont I, Borracci P, Moukha S, Carratu MR. Synergistic effects of fumonisin B1 and ochratoxin A: are in vitro cytotoxicity data predictive of in vivo acute toxicity? Toxicology. 2004;201(1-3):115-23. 258. McLaughlin J, Padfield PJ, Burt JP, O'Neill CA. Ochratoxin A increases permeability through tight junctions by removal of specific claudin isoforms. American journal of physiology Cell physiology. 2004;287(5):C1412-7. 259. Mobio TA, Tavan E, Baudrimont I, Anane R, Carratú M-R, Sanni A, et al. Comparative study of the toxic effects of fumonisin B1 in rat C6 glioma cells and p53-null mouse embryo fibroblasts. Toxicology. 2003;183(1):65-75. 260. Kouadio JH, Dano SD, Moukha S, Mobio TA, Creppy EE. Effects of combinations of Fusarium mycotoxins on the inhibition of macromolecular synthesis, malondialdehyde levels, DNA methylation and fragmentation, and viability in Caco-2 cells. Toxicon : official journal of the International Society on Toxinology. 2007;49(3):306-17. 261. De Walle JV, Sergent T, Piront N, Toussaint O, Schneider YJ, Larondelle Y. Deoxynivalenol affects in vitro intestinal epithelial cell barrier integrity through inhibition of protein synthesis. Toxicology and applied pharmacology. 2010;245(3):291-8. 262. Ostry V, Malir F, Toman J, Grosse Y. Mycotoxins as human carcinogens-the IARC Monographs classification. Mycotoxin Res. 2017;33(1):65-73. 263. Gerez JR, Pinton P, Callu P, Grosjean F, Oswald IP, Bracarense AP. Deoxynivalenol alone or in combination with nivalenol and zearalenone induce systemic histological changes in pigs. Experimental and toxicologic pathology : official journal of the Gesellschaft fur Toxikologische Pathologie. 2015;67(2):89-98. 264. Liang Z, Ren Z, Gao S, Chen Y, Yang Y, Yang D, et al. Individual and combined effects of deoxynivalenol and zearalenone on mouse kidney. Environmental toxicology and pharmacology. 2015;40(3):686-91. 265. Zhou H, George S, Hay C, Lee J, Qian H, Sun X. Individual and combined effects of Aflatoxin B1, Deoxynivalenol and Zearalenone on HepG2 and RAW 264.7 cell lines. Food and chemical toxicology : an international journal published for the British Industrial Biological Research Association. 2017;103:18-27. 266. Akbari P, Braber S, Varasteh S, Alizadeh A, Garssen J, Fink-Gremmels J. The intestinal barrier as an emerging target in the toxicological assessment of mycotoxins. Archives of toxicology. 2017;91(3):1007-29. 267. Assuncao R, Ferreira M, Martins C, Diaz I, Padilla B, Dupont D, et al. Applicability of in vitro methods to study patulin bioaccessibility and its effects on intestinal membrane integrity. J Toxicol Environ Health A. 2014;77(14-16):983-92. 268. Rodríguez-Carrasco Y, Ruiz MJ, Font G, Berrada H. Exposure estimates to Fusarium mycotoxins through cereals intake. Chemosphere. 2013;93(10):2297-303. 269. Smith M-C, Madec S, Coton E, Hymery N. Natural Co-Occurrence of Mycotoxins in Foods and Feeds and Their in vitro Combined Toxicological Effects. Toxins. 2016;8(4):94. 270. Manda G, Mocanu MA, Marin DE, Taranu I. Dual effects exerted in vitro by micromolar concentrations of deoxynivalenol on undifferentiated caco-2 cells. Toxins. 2015;7(2):593-603. 271. Cano-Sancho G, Gonzalez-Arias CA, Ramos AJ, Sanchis V, Fernandez-Cruz ML. Cytotoxicity of the mycotoxins deoxynivalenol and ochratoxin A on Caco-2 cell line in

186

REFERENCES

presence of resveratrol. Toxicology in vitro : an international journal published in association with BIBRA. 2015;29(7):1639-46. 272. Juan-Garcia A, Juan C, Konig S, Ruiz MJ. Cytotoxic effects and degradation products of three mycotoxins: alternariol, 3-acetyl-deoxynivalenol and 15-acetyldeoxynivalenol in liver hepatocellular carcinoma cells. Toxicology letters. 2015;235(1):8-16. 273. Alassane-Kpembi I, Kolf-Clauw M, Gauthier T, Abrami R, Abiola FA, Oswald IP, et al. New insights into mycotoxin mixtures: the toxicity of low doses of Type B trichothecenes on intestinal epithelial cells is synergistic. Toxicology and applied pharmacology. 2013;272(1):191-8. 274. Alassane-Kpembi I, Puel O, Oswald IP. Toxicological interactions between the mycotoxins deoxynivalenol, nivalenol and their acetylated derivatives in intestinal epithelial cells. Archives of toxicology. 2015;89(8):1337-46. 275. Alassane-Kpembi I, Schatzmayr G, Taranu I, Marin D, Puel O, Oswald IP. Mycotoxins co-contamination: Methodological aspects and biological relevance of combined toxicity studies. Critical reviews in food science and nutrition. 2017;57(16):3489-507. 276. Chou TC. Drug combination studies and their synergy quantification using the Chou- Talalay method. Cancer research. 2010;70(2):440-6. 277. Chou TC. Theoretical basis, experimental design, and computerized simulation of synergism and antagonism in drug combination studies. Pharmacological reviews. 2006;58(3):621-81. 278. Luo S, Terciolo C, Bracarense A, Payros D, Pinton P, Oswald IP. In vitro and in vivo effects of a mycotoxin, deoxynivalenol, and a trace metal, cadmium, alone or in a mixture on the intestinal barrier. Environ Int. 2019;132:105082. 279. Lee JG, Kim SY, Moon JS, Kim SH, Kang DH, Yoon HJ. Effects of grilling procedures on levels of polycyclic aromatic hydrocarbons in grilled meats. Food Chem. 2016;199:632-8. 280. Kumar Y, Yadav DN, Ahmad T, Narsaiah K. Recent Trends in the Use of Natural Antioxidants for Meat and Meat Products. Comprehensive Reviews in Food Science and Food Safety. 2015;14(6):796-812. 281. Duthie G, Campbell F, Bestwick C, Stephen S, Russell W. Antioxidant effectiveness of vegetable powders on the lipid and protein oxidative stability of cooked Turkey meat patties: implications for health. Nutrients. 2013;5(4):1241-52. 282. Emami A, Nasri MH, Ganjkhanlou M, Zali A, Rashidi L. Effects of dietary pomegranate seed pulp on oxidative stability of kid meat. Meat science. 2015;104:14-9. 283. Gallo M, Ferracane R, Naviglio D. Antioxidant addition to prevent lipid and protein oxidation in chicken meat mixed with supercritical extracts of Echinacea angustifolia. The Journal of Supercritical Fluids. 2012;72:198-204. 284. Kim S-J, Cho AR, Han J. Antioxidant and antimicrobial activities of leafy green vegetable extracts and their applications to meat product preservation. Food Control. 2013;29(1):112-20. 285. Price A, Diaz P, Banon S, Garrido MD. Natural extracts versus sodium ascorbate to extend the shelf life of meat-based ready-to-eat meals. Food science and technology international = Ciencia y tecnologia de los alimentos internacional. 2013;19(5):427- 38. 286. Selani MM, Contreras-Castillo CJ, Shirahigue LD, Gallo CR, Plata-Oviedo M, Montes- Villanueva ND. Wine industry residues extracts as natural antioxidants in raw and cooked chicken meat during frozen storage. Meat science. 2011;88(3):397-403. 287. Haskaraca G, Demirok E, Kolsarıcı N, Öz F, Özsaraç N. Effect of green tea extract and microwave pre-cooking on the formation of heterocyclic aromatic amines in fried chicken meat products. Food research international (Ottawa, Ont). 2014;63:373-81. 288. Janoszka B. Heterocyclic amines and azaarenes in pan-fried meat and its gravy fried without additives and in the presence of onion and garlic. Food Chem. 2010;120(2):463-73.

187

REFERENCES

289. Keskekoglu H, Uren A. Inhibitory effects of pomegranate seed extract on the formation of heterocyclic aromatic amines in beef and chicken meatballs after cooking by four different methods. Meat science. 2014;96(4):1446-51. 290. Van Hecke T, Wouters A, Rombouts C, Izzati T, Berardo A, Vossen E, et al. Reducing Compounds Equivocally Influence Oxidation during Digestion of a High-Fat Beef Product, which Promotes Cytotoxicity in Colorectal Carcinoma Cell Lines. Journal of agricultural and food chemistry. 2016;64(7):1600-9. 291. Li, Henning SM, Zhang Y, Rahnama N, Zerlin A, Thames G, et al. Decrease of postprandial endothelial dysfunction by spice mix added to high-fat hamburger meat in men with Type 2 diabetes mellitus. Diabetic medicine : a journal of the British Diabetic Association. 2013;30(5):590-5. 292. Li, Henning SM, Zhang Y, Zerlin A, Li L, Gao K, et al. Antioxidant-rich spice added to hamburger meat during cooking results in reduced meat, plasma, and urine malondialdehyde concentrations. Am J Clin Nutr. 2010;91(5):1180-4. 293. Busquets R, Puignou L, Galceran MT, Skog K. Effect of Red Wine Marinades on the Formation of Heterocyclic Amines in Fried Chicken Breast. J Agric Food Chem. 2006;54(21):8376-84. 294. Melo A, Viegas O, Petisca C, Pinho O, Ferreira IMPLVO. Effect of Beer/Red Wine Marinades on the Formation of Heterocyclic Aromatic Amines in Pan-Fried Beef. Journal of Agricultural and Food Chemistry. 2008;56(22):10625-32. 295. Viegas O, Moreira PS, Ferreira IM. Influence of beer marinades on the reduction of carcinogenic heterocyclic aromatic amines in charcoal-grilled pork meat. Food additives & contaminants Part A, Chemistry, analysis, control, exposure & risk assessment. 2015;32(3):315-23. 296. Gorelik, Ligumsky M, Kohen R, Kanner J. A novel function of red wine polyphenols in humans: prevention of absorption of cytotoxic lipid peroxidation products. FASEB journal : official publication of the Federation of American Societies for Experimental Biology. 2008;22(1):41-6. 297. Tagliazucchi D, Verzelloni E, Conte A. Effect of Dietary Melanoidins on Lipid Peroxidation during Simulated Gastric Digestion: Their Possible Role in the Prevention of Oxidative Damage. Journal of Agricultural and Food Chemistry. 2010;58(4):2513-9. 298. Homem V, Santos L. Degradation and removal methods of antibiotics from aqueous matrices--a review. Journal of environmental management. 2011;92(10):2304-47. 299. Bushra R, Aslam N, Khan AY. Food-drug interactions. Oman medical journal. 2011;26(2):77-83. 300. Welling PG. Interactions affecting drug absorption. Clinical pharmacokinetics. 1984;9(5):404-34. 301. Delgado J, Núñez F, Rodríguez M. Mycotoxins in meat and meat products. Especial Micotoxins. Available at: https://issuu.com/grupoagrinews/docs/mycotoxins-in- meat?e=26014920/68087128 (last accessed 12/12/2019). 2019. 302. Panda P, Mehta A. Aflatoxin Detoxification Potential of Ocimum Tenuiflorum. Journal of Food Safety. 2013;33:265-72. 303. Vijayanandraj S, Brinda R, Kannan K, Adhithya R, Vinothini S, Senthil K, et al. Detoxification of aflatoxin B1 by an aqueous extract from leaves of Adhatoda vasica Nees. Microbiological Research. 2014;169(4):294-300. 304. Sultana B, Naseer R, Nigam P. Utilization of agro-wastes to inhibit aflatoxins synthesis by Aspergillus parasiticus: A biotreatment of three cereals for safe long-term storage. Bioresource Technology. 2015;197:443-50. 305. Udomkun P, Wiredu AN, Nagle M, Müller J, Vanlauwe B, Bandyopadhyay R. Innovative technologies to manage aflatoxins in foods and feeds and the profitability of application – A review. Food Control. 2017;76:127-38. 306. Velazhahan R, Vijayanandraj S, Vijayasamundeeswari A, Paranidharan V, Samiyappan R, Iwamoto T, et al. Detoxification of aflatoxins by seed extracts of the

188

REFERENCES

medicinal plant, Trachyspermum ammi (L.) Sprague ex Turrill – Structural analysis and biological toxicity of degradation product of aflatoxin G1. Food Control. 2010;21(5):719-25. 307. Oliveira H, Muniz JA, Bandarra NM, Castanheira I, Coelho IR, Delgado I, et al. Effects of Industrial Boiling on the Nutritional Profile of Common Octopus (Octopus vulgaris). Foods (Basel, Switzerland). 2019;8(9). 308. Egger L, Ménard O, Delgado-Andrade C, Alvito P, Assunção R, Balance S, et al. The harmonized INFOGEST in vitro digestion method: From knowledge to action. Food Research International. 2016;88:217-25. 309. EC. European Commision. No 1881/2006 of 19 December 2006 setting maximum levels for certain contaminants in foodstuffs ([Last accessed 14.05.2019) https://eur- lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:02006R1881- 20100701&from=EN. 2006. 310. AOAC. Association of Oﬃcial Analytical Chemists International. Oﬃcial Methods of Analysis, 16th ed.; AOAC International: Washington, DC, USA, 1998. . 1998. 311. Dashdorj D, Hwang I. Effect of extraction methods on the types and levels of free amino acid of beef longissimus muscle. Korean J Food Sci Anim Resour 2012;32(369- 375). 312. Badawy AA. The EZ:Faast family of amino acid analysis kits: application of the GC- FID kit for rapid determination of plasma tryptophan and other amino acids. Methods in molecular biology (Clifton, NJ). 2012;828:153-64. 313. Cruz R, Casal S, Mendes E, Costa A, Santos C, Morais S. Validation of a Single- Extraction Procedure for Sequential Analysis of Vitamin E, Cholesterol, Fatty Acids, and Total Fat in Seafood. Food Analytical Methods. 2013;6(4):1196-204. 314. Dalziel CJ, Kliem KE, Givens DI. Fat and fatty acid composition of cooked meat from UK retail chickens labelled as from organic and non-organic production systems. Food Chemistry. 2015;179:103-8. 315. Mendes R, Cardoso C, Pestana C. Measurement of malondialdehyde in fish: A comparison study between HPLC methods and the traditional spectrophotometric test. Food Chemistry. 2009;112(4):1038-45. 316. Mezzar S, de Schryver E, Van Veldhoven PP. RP-HPLC-fluorescence analysis of aliphatic aldehydes: application to aldehyde-generating enzymes HACL1 and SGPL1. Journal of lipid research. 2014;55(3):573-82. 317. Ganhao R, Morcuende D, Estevez M. Protein oxidation in emulsified cooked burger patties with added fruit extracts: Influence on colour and texture deterioration during chill storage. Meat science. 2010;85(3):402-9. 318. Hu, Ren S, Shen Q, Ye X, Chen J, Ling J. Protein oxidation and proteolysis during roasting and in vitro digestion of fish (Acipenser gueldenstaedtii). Journal of the science of food and agriculture. 2018;98(14):5344-51. 319. Bittencourt MS, Martins MT, de Albuquerque FG, Barreto F, Hoff R. High-throughput multiclass screening method for antibiotic residue analysis in meat using liquid chromatography-tandem mass spectrometry: a novel minimum sample preparation procedure. Food additives & contaminants Part A, Chemistry, analysis, control, exposure & risk assessment. 2012;29(4):508-16. 320. Moretti S, Dusi G, Giusepponi D, Pellicciotti S, Rossi R, Saluti G, et al. Screening and confirmatory method for multiclass determination of 62 antibiotics in meat. Journal of Chromatography A. 2016;1429:175-88. 321. Palm GJ, Lederer T, Orth P, Saenger W, Takahashi M, Hillen W, et al. Specific binding of divalent metal ions to tetracycline and to the Tet repressor/tetracycline complex. Journal of biological inorganic chemistry : JBIC : a publication of the Society of Biological Inorganic Chemistry. 2008;13(7):1097-110. 322. Gomez-Perez ML, Plaza-Bolanos P, Romero-Gonzalez R, Martinez-Vidal JL, Garrido-Frenich A. Comprehensive qualitative and quantitative determination of

189

REFERENCES

pesticides and veterinary drugs in honey using liquid chromatography-Orbitrap high resolution mass spectrometry. Journal of chromatography A. 2012;1248:130-8. 323. Yogendrarajah P, Van Poucke C, De Meulenaer B, De Saeger S. Development and validation of a QuEChERS based liquid chromatography tandem mass spectrometry method for the determination of multiple mycotoxins in spices. Journal of chromatography A. 2013;1297:1-11. 324. Cunha SC, Sa SVM, Fernandes JO. Multiple mycotoxin analysis in nut products: Occurrence and risk characterization. Food and chemical toxicology : an international journal published for the British Industrial Biological Research Association. 2018;114:260-9. 325. Commission. Commission Decision of 12 August 2002 implementing Council Directive 96/23/EC concerning the performance of analytical methods and the interpretation of results. Official Journal of the European Communities. 2002/657/EC. 2002. 326. ICH. IHC harmonized tripartite guideline. Validation of analytical procedures: text and methodology Q2(R1). 2005. 327. SANTE. Method validation and analytical quality control procedures for pesticide residues analysis in food & feed. SANTE/12682/2019. 2019. 328. Chen S, Einspanier R, Schoen J. Transepithelial electrical resistance (TEER): a functional parameter to monitor the quality of oviduct epithelial cells cultured on filter supports. Histochem Cell Biol. 2015;144(5):509-15. 329. Tavelin S, Grasjo J, Taipalensuu J, Ocklind G, Artursson P. Applications of epithelial cell culture in studies of drug transport. Methods in molecular biology (Clifton, NJ). 2002;188:233-72. 330. Skolnik S, Lin X, Wang J, Chen XH, He T, Zhang B. Towards prediction of in vivo intestinal absorption using a 96-well Caco-2 assay. Journal of pharmaceutical sciences. 2010;99(7):3246-65. 331. Tavelin S, Taipalensuu J, Söderberg L, Morrison R, Chong S, Artursson P. Prediction of the Oral Absorption of Low-Permeability Drugs Using Small Intestine-Like 2/4/A1 Cell Monolayers. Pharmaceutical Research. 2003;20(3):397-405. 332. Chou T-C, Talalay P. Quantitative analysis of dose-effect relationships: the combined effects of multiple drugs or enzyme inhibitors. Advances in Enzyme Regulation. 1984;22(Supplement C):27-55. 333. Sarstedt M, Mooi E. Cluster analysis. A Concise Guide to Market Research. STBE, pp. 273-324. Springer, Heidelberg. doi: 10.1007/978-3-642-53965-9. 2014. 334. Shahidi F, Zhong Y. Lipid oxidation and improving the oxidative stability. Chemical Society reviews. 2010;39(11):4067-79. 335. Van Hecke T, Goethals S, Vossen E, De Smet S. Long-Chain n-3 PUFA Content and n-6/n-3 PUFA Ratio in Mammal, Poultry, and Fish Muscles Largely Explain Differential Protein and Lipid Oxidation Profiles Following In Vitro Gastrointestinal Digestion. Molecular nutrition & food research. 2019;63(22):e1900404. 336. Aron PM, Shellhammer TH. A Discussion of Polyphenols in Beer Physical and Flavour Stability. Journal of the Institute of Brewing. 2010;116(4):369-80. 337. Das SK, Vasudevan DM. Alcohol-induced oxidative stress. Life Sciences. 2007;81(3):177-87. 338. Gutierrez-Grijalva EP, Picos-Salas MA, Leyva-Lopez N, Criollo-Mendoza MS, Vazquez-Olivo G, Heredia JB. Flavonoids and Phenolic Acids from Oregano: Occurrence, Biological Activity and Health Benefits. Plants (Basel, Switzerland). 2017;7(1). 339. Amaral AB, Silva MVd, Lannes SCdS. Lipid oxidation in meat: mechanisms and protective factors - a review. Food Science and Technology. 2018;38:1-15. 340. Weber D, Davies MJ, Grune T. Determination of protein carbonyls in plasma, cell extracts, tissue homogenates, isolated proteins: Focus on sample preparation and derivatization conditions. Redox biology. 2015;5:367-80.

190

REFERENCES

341. Han JH, Lee SM, Kim YS. Effects of Schiff Base Formation and Aldol Condensation on the Determination of Aldehydes in Rice Wine Using GC-MS. Molecules. 2017;22(4). 342. Veberg A, Vogt G, Wold JP. Fluorescence in aldehyde model systems related to lipid oxidation. LWT - Food Science and Technology. 2006;39(5):562-70. 343. Hur SJ, Lee SY, Moon SS, Lee SJ. In Vitro Effects of Cooking Methods on Digestibility of Lipids and Formation of Cholesterol Oxidation Products in Pork. Korean J Food Sci Anim Resour. 2014;34(3):280-6. 344. Raes K, Doolaege EHA, Deman S, Vossen E, De Smet S. Effect of carnosic acid, quercetin and α-tocopherol on lipid and protein oxidation in an in vitro simulated gastric digestion model. International Journal of Food Sciences and Nutrition. 2015;66(2):216-21. 345. Gorelik, Lapidot T, Shaham I, Granit R, Ligumsky M, Kohen R, et al. Lipid Peroxidation and Coupled Vitamin Oxidation in Simulated and Human Gastric Fluid Inhibited by Dietary Polyphenols: Health Implications. Journal of Agricultural and Food Chemistry. 2005;53(9):3397-402. 346. Kaur L, Maudens E, Haisman DR, Boland MJ, Singh H. Microstructure and protein digestibility of beef: The effect of cooking conditions as used in stews and curries. LWT - Food Science and Technology. 2014;55(2):612-20. 347. Wen S, Zhou G, Song S, Xu X, Voglmeir J, Liu L, et al. Discrimination of in vitro and in vivo digestion products of meat proteins from pork, beef, chicken, and fish. Proteomics. 2015;15(21):3688-98. 348. Stolker AA, Brinkman UA. Analytical strategies for residue analysis of veterinary drugs and growth-promoting agents in food-producing animals--a review. Journal of chromatography A. 2005;1067(1-2):15-53. 349. Deschamps E, Vasconcelos O, Lange L, Donnici CL, Silva MCd, Sales JA. Management of effluents and waste from pharmaceutical industry in Minas Gerais, Brazil. Brazilian Journal of Pharmaceutical Sciences. 2012;48:727-36. 350. Roca M, Castillo M, Marti P, Althaus RL, Molina MP. Effect of heating on the stability of quinolones in milk. Journal of agricultural and food chemistry. 2010;58(9):5427-31. 351. Bobu M, Yediler A, Siminiceanu I, Zhang F, Schulte-Hostede S. Comparison of different advanced oxidation processes for the degradation of two fluoroquinolone antibiotics in aqueous solutions. J Environ Sci Health A Tox Hazard Subst Environ Eng. 2013;48(3):251-62. 352. Zhang H, Huang C-H. Oxidative Transformation of Fluoroquinolone Antibacterial Agents and Structurally Related Amines by Manganese Oxide. Environmental science & technology. 2005;39(12):4474-83. 353. Wenk J, Canonica S. Phenolic Antioxidants Inhibit the Triplet-Induced Transformation of Anilines and Sulfonamide Antibiotics in Aqueous Solution. Environmental science & technology. 2012;46(10):5455-62. 354. Heshmati A. Impact of Cooking Procedures on Antibacterial Drug Residues in Food: A Review. Journal of Food Qualirt and Hazards Control. 2015;2:33-7. 355. Hussein M, Ahmed M, Morshedy A. Effect of cooking methods on some antibiotic residues in chicken meat. Jpn J Vet Res 2016;64 (Supplement 2):S225-S31. 356. Ibrahim A, Moats WA. Effect of Cooking Procedures on Oxytetracycline Residues in Lamb Muscle. Journal of Agricultural and Food Chemistry. 1994;42(11):2561-3. 357. Rose MD, Bygrave J, Farrington WHH, Shearer G. The effect of cooking on veterinary drug residues in food: 4. Oxytetracycline. Food Additives & Contaminants. 1996;13(3):275-86. 358. Ismail-Fitry MR, Jinap S, Jamilah B, Saleha AA. Effect of Different Time and Temperature of Various Cooking Methods on Sulfonamide Residues in Chicken Balls. In: Gökçekus H, Türker U, LaMoreaux JW, editors. Survival and Sustainability: Environmental concerns in the 21st Century. Berlin, Heidelberg: Springer Berlin Heidelberg; 2011. p. 607-13.

191

REFERENCES

359. Furusawa N, Hanabusa R. Cooking effects on sulfonamide residues in chicken thigh muscle. Food Research International. 2002;35(1):37-42. 360. McEneff G, Barron L, Kelleher B, Paull B, Quinn B. The determination of pharmaceutical residues in cooked and uncooked marine bivalves using pressurised liquid extraction, solid-phase extraction and liquid chromatography-tandem mass spectrometry. Analytical and bioanalytical chemistry. 2013;405(29):9509-21. 361. Bacila DM, Cunha A, Jr., Gressler V, Scheuermann GN, Coldebella A, Caron L, et al. Detection of p-Nitroaniline Released from Degradation of 4,4'-Dinitrocarbanilide in Chicken Breast during Thermal Processing. Journal of agricultural and food chemistry. 2019;67(32):9002-8. 362. Koziolek M, Alcaro S, Augustijns P, Basit AW, Grimm M, Hens B, et al. The mechanisms of pharmacokinetic food-drug interactions – A perspective from the UNGAP group. European Journal of Pharmaceutical Sciences. 2019;134:31-59. 363. McLachlan A, Ramzan I. Meals and medicines. Aust Prescr. 2006;29:40-2. . 364. Mukherjee M, Sardar PS, Ghorai SK, Samanta SK, Roy AS, Dasgupta S, et al. A comparative study of interaction of tetracycline with several proteins using time resolved anisotropy, phosphorescence, docking and FRET. PloS one. 2013;8(4):e60940. 365. Nielsen P, Gyrd-Hansen N. Bioavailability of oxytetracycline, tetracycline and chlortetracycline after oral administration to fed and fasted pigs. Journal of veterinary pharmacology and therapeutics. 1996;19(4):305-11. 366. de Lastours V, Fantin B. Impact of fluoroquinolones on human microbiota. Focus on the emergence of antibiotic resistance. Future microbiology. 2015;10(7):1241-55. 367. Jung JY, Ahn Y, Khare S, Gokulan K, Pineiro SA, Cerniglia CE. An in vitro study to assess the impact of tetracycline on the human intestinal microbiome. Anaerobe. 2018;49:85-94. 368. Marker LM, Hammer AS, Andresen L, Isaack P, Clausen T, Byskov K, et al. Short- term effect of oral amoxicillin treatment on the gut microbial community composition in farm mink (Neovison vison). FEMS Microbiology Ecology. 2017;93(7). 369. Wolf-Hall CE, Hanna MA, Bullerman LB. Stability of deoxynivalenol in heat-treated foods. Journal of food protection. 1999;62(8):962-4. 370. Bretz M, Beyer M, Cramer B, Knecht A, Humpf HU. Thermal degradation of the Fusarium mycotoxin deoxynivalenol. Journal of agricultural and food chemistry. 2006;54(17):6445-51. 371. Raters M, Matissek R. Thermal stability of aflatoxin B1 and ochratoxin A. Mycotoxin Res. 2008;24(3):130-4. 372. Seefelder W, Knecht A, Humpf HU. Bound fumonisin B1: analysis of fumonisin-B1 glyco and amino acid conjugates by liquid chromatography-electrospray ionization- tandem mass spectrometry. Journal of agricultural and food chemistry. 2003;51(18):5567-73. 373. Maragos C. Zearalenone occurrence and human exposure. World Mycotoxin Journal. 2010;3(4):369-83. 374. Ryu D, Hanna MA, Eskridge KM, Bullerman LB. Heat stability of zearalenone in an aqueous buffered model system. Journal of agricultural and food chemistry. 2003;51(6):1746-8. 375. Embuscado ME. Spices and herbs: Natural sources of antioxidants – a mini review. Journal of Functional Foods. 2015;18:811-9. 376. Peng Z, Chen L, Zhu Y, Huang Y, Hu X, Wu Q, et al. Current major degradation methods for aflatoxins: A review. Trends in Food Science & Technology. 2018;80:155-66. 377. Hajare SS, Hajare SN, Sharma A. Aflatoxin Inactivation Using Aqueous Extract of Ajowan (Trachyspermum ammi) Seeds. Journal of Food Science. 2005;70(1):C29- C34.

192

REFERENCES

378. Pragyanshree P, Alka M. Aflatoxin detoxification potential of ocimum tenuiflorum. Journal of Food Safety. 2013;33(3):265-72. 379. Vijayanandraj S, Brinda R, Kannan K, Adhithya R, Vinothini S, Senthil K, et al. Detoxification of aflatoxin B1 by an aqueous extract from leaves of Adhatoda vasica Nees. Microbiological research. 2014;169(4):294-300. 380. Mallebrera B, Meca G, Manyes L, Mañes J, Font G. Influence of pro- and prebiotics on gastric, duodenal and colonic bioaccessibility of the mycotoxin beauvericin. Journal of Food Composition and Analysis. 2013;32(2):141-9. 381. Melo A, Faria MA, Pinto E, Mansilha C, Ferreira I. In vitro bioacessibility and transport across Caco-2 monolayers of haloacetic acids in drinking water. Chemosphere. 2016;161:19-26. 382. Hurni MA, Noach AB, Blom-Roosemalen MC, de Boer AG, Nagelkerke JF, Breimer DD. Permeability enhancement in Caco-2 cell monolayers by sodium salicylate and sodium taurodihydrofusidate: assessment of effect-reversibility and imaging of transepithelial transport routes by confocal laser scanning microscopy. The Journal of pharmacology and experimental therapeutics. 1993;267(2):942-50. 383. Mahfoud R, Maresca M, Santelli M, Pfohl-Leszkowicz A, Puigserver A, Fantini J. pH- dependent interaction of fumonisin B1 with cholesterol: physicochemical and molecular modeling studies at the air-water interface. Journal of agricultural and food chemistry. 2002;50(2):327-31. 384. Tomkova I, Sevcikova Z, Levkut M, Revajova V, Conkova E, Laciakova A, et al. Effect of aflatoxin B1 on CD3 T cells and alkaline phosphatase in the intestine of mice. Mycopathologia. 2002;154(1):15-9. 385. Artursson P, Palm K, Luthman K. Caco-2 monolayers in experimental and theoretical predictions of drug transport. Adv Drug Deliv Rev. 2001;46(1-3):27-43. 386. Yu LX, Lipka E, Crison JR, Amidon GL. Transport approaches to the biopharmaceutical design of oral drug delivery systems: prediction of intestinal absorption. Adv Drug Deliv Rev. 1996;19(3):359-76. 387. Antonissen G, Devreese M, Van Immerseel F, De Baere S, Hessenberger S, Martel A, et al. Chronic exposure to deoxynivalenol has no influence on the oral bioavailability of fumonisin B1 in broiler chickens. Toxins. 2015;7(2):560-71. 388. Nistor A-M, Cotan Ş-D, Nechita C-B, Tarţian A, Niculaua M, Cotea VV. Rapid assessment of mycotoxins in wine by on-line SPE-UHPLC-FLD. BIO Web Conf. 2017;9. 389. Afsah-Hejri L, Jinap S, Hajeb P, Radu S, Shakibazadeh S. A Review on Mycotoxins in Food and Feed: Malaysia Case Study. Comprehensive Reviews in Food Science and Food Safety. 2013;12(6):629-51. 390. Qi X, Wagenaar E, Xu W, Huang K, Schinkel AH. Ochratoxin A transport by the human breast cancer resistance protein (BCRP), multidrug resistance protein 2 (MRP2), and organic anion-transporting polypeptides 1A2, 1B1 and 2B1. Toxicology and applied pharmacology. 2017;329:18-25. 391. Schrickx J, Lektarau Y, Fink-Gremmels J. Ochratoxin A secretion by ATP-dependent membrane transporters in Caco-2 cells. Archives of toxicology. 2006;80(5):243-9. 392. Sergent T, Ribonnet L, Kolosova A, Garsou S, Schaut A, De Saeger S, et al. Molecular and cellular effects of food contaminants and secondary plant components and their plausible interactions at the intestinal level. Food and chemical toxicology : an international journal published for the British Industrial Biological Research Association. 2008;46(3):813-41. 393. Gratz S, Wu QK, El-Nezami H, Juvonen RO, Mykkanen H, Turner PC. Lactobacillus rhamnosus strain GG reduces aflatoxin B1 transport, metabolism, and toxicity in Caco-2 Cells. Applied and environmental microbiology. 2007;73(12):3958-64. 394. Mata JE, Yu Z, Gray JE, Williams DE, Rodriguez-Proteau R. Effects of chlorophyllin on transport of dibenzo(a, l)pyrene, 2-amino-1-methyl-6-phenylimidazo-[4,5-

193

REFERENCES

b]pyridine, and aflatoxin B(1) across Caco-2 cell monolayers. Toxicology. 2004;196(1- 2):117-25. 395. Tachampa K, Takeda M, Khamdang S, Noshiro-Kofuji R, Tsuda M, Jariyawat S, et al. Interactions of organic anion transporters and organic cation transporters with mycotoxins. Journal of pharmacological sciences. 2008;106(3):435-43. 396. Volk C. OCTs, OATs, and OCTNs: structure and function of the polyspecific organic ion transporters of the SLC22 family. Wiley Interdisciplinary Reviews: Membrane Transport and Signaling. 2014;3(1):1-13. 397. Loe DW, Stewart RK, Massey TE, Deeley RG, Cole SP. ATP-dependent transport of aflatoxin B1 and its glutathione conjugates by the product of the multidrug resistance protein (MRP) gene. Molecular pharmacology. 1997;51(6):1034-41. 398. Gallo A, Moschini M, Masoero F. Aflatoxins absorption in the gastro-intestinal tract and in the vaginal mucosa in lactating dairy cows. Italian Journal of Animal Science. 2008;7(1):53-63. 399. Jubert C, Mata J, Bench G, Dashwood R, Pereira C, Tracewell W, et al. Effects of chlorophyll and chlorophyllin on low-dose aflatoxin B(1) pharmacokinetics in human volunteers. Cancer prevention research (Philadelphia, Pa). 2009;2(12):1015-22. 400. Knutsen HK, Alexander J, Barregård L, Bignami M, Brüschweiler B, Ceccatelli S, et al. Risks to human and animal health related to the presence of deoxynivalenol and its acetylated and modified forms in food and feed. EFSA Journal. 2017;15(9):e04718. 401. Lee HJ, Ryu D. Worldwide Occurrence of Mycotoxins in Cereals and Cereal-Derived Food Products: Public Health Perspectives of Their Co-occurrence. Journal of Agricultural and Food Chemistry. 2017;65(33):7034-51. 402. FDA. Drug Interaction Studies – Study Design, Data Analysis, and Implications for Dosing and Labeling. . 2006. 403. Sergent T, Garsou S, Schaut A, De Saeger S, Pussemier L, Van Peteghem C, et al. Differential modulation of ochratoxin A absorption across Caco-2 cells by dietary polyphenols, used at realistic intestinal concentrations. Toxicology letters. 2005;159(1):60-70. 404. Lusky K, Tesch D, Gobel R, Haider W. GleichzeitigeVerabreichung der Mykotoxine Ochratoxin A und Zearalenone u¨ber das Futter an Schweine – Einfluss auf Tiergesundheit und Ru¨ckstandsverhalten. Tiera¨rztliche Umschau. 1998;52:212-21. 405. Clarke R, Connolly L, Frizzell C, Elliott CT. Cytotoxic assessment of the regulated, co-existing mycotoxins aflatoxin B1, fumonisin B1 and ochratoxin, in single, binary and tertiary mixtures. Toxicon : official journal of the International Society on Toxinology. 2014;90:70-81. 406. Hundhausen C, Bosch-Saadatmandi C, Augustin K, Blank R, Wolffram S, Rimbach G. Effect of vitamin E and polyphenols on ochratoxin A-induced cytotoxicity in liver (HepG2) cells. Journal of plant physiology. 2005;162(7):818-22. 407. McKean C, Tang L, Billam M, Tang M, Theodorakis CW, Kendall RJ, et al. Comparative acute and combinative toxicity of aflatoxin B1 and T-2 toxin in animals and immortalized human cell lines. Journal of applied toxicology : JAT. 2006;26(2):139-47. 408. Chuturgoon AA, Phulukdaree A, Moodley D. Fumonisin B1 inhibits apoptosis in HepG2 cells by inducing Birc-8/ILP-2. Toxicology Letters. 2015;235(2):67-74. 409. Sun LH, Lei MY, Zhang NY, Gao X, Li C, Krumm CS, et al. Individual and combined cytotoxic effects of aflatoxin B1, zearalenone, deoxynivalenol and fumonisin B1 on BRL 3A rat liver cells. Toxicon : official journal of the International Society on Toxinology. 2015;95:6-12. 410. Wan LY, Turner PC, El-Nezami H. Individual and combined cytotoxic effects of Fusarium toxins (deoxynivalenol, nivalenol, zearalenone and fumonisins B1) on swine jejunal epithelial cells. Food and chemical toxicology : an international journal published for the British Industrial Biological Research Association. 2013;57:276-83.

194

REFERENCES

411. Arunachalam C, Doohan FM. Trichothecene toxicity in eukaryotes: cellular and molecular mechanisms in plants and animals. Toxicology letters. 2013;217(2):149- 58. 412. Alizadeh A, Braber S, Akbari P, Garssen J, Fink-Gremmels J. Deoxynivalenol Impairs Weight Gain and Affects Markers of Gut Health after Low-Dose, Short-Term Exposure of Growing Pigs. Toxins. 2015;7(6):2071. 413. Bracarense AP, Lucioli J, Grenier B, Drociunas Pacheco G, Moll WD, Schatzmayr G, et al. Chronic ingestion of deoxynivalenol and fumonisin, alone or in interaction, induces morphological and immunological changes in the intestine of piglets. The British journal of nutrition. 2012;107(12):1776-86. 414. Nossol C, Diesing AK, Kahlert S, Kersten S, Kluess J, Ponsuksili S, et al. Deoxynivalenol affects the composition of the basement membrane proteins and influences en route the migration of CD16(+) cells into the intestinal epithelium. Mycotoxin Research. 2013;29(4):245-54. 415. Pinton P, Tsybulskyy D, Lucioli J, Laffitte J, Callu P, Lyazhri F, et al. Toxicity of deoxynivalenol and its acetylated derivatives on the intestine: differential effects on morphology, barrier function, tight junction proteins, and mitogen-activated protein kinases. Toxicological sciences : an official journal of the Society of Toxicology. 2012;130(1):180-90. 416. Wu M, Xiao H, Ren W, Yin J, Tan B, Liu G, et al. Therapeutic Effects of Glutamic Acid in Piglets Challenged with Deoxynivalenol. PloS one. 2014;9(7):e100591. 417. Xiao H, Tan BE, Wu MM, Yin YL, Li TJ, Yuan DX, et al. Effects of composite antimicrobial peptides in weanling piglets challenged with deoxynivalenol: II. Intestinal morphology and function. Journal of animal science. 2013;91(10):4750-6. 418. Lucioli J, Pinton P, Callu P, Laffitte J, Grosjean F, Kolf-Clauw M, et al. The food contaminant deoxynivalenol activates the mitogen activated protein kinases in the intestine: interest of ex vivo models as an alternative to in vivo experiments. Toxicon : official journal of the International Society on Toxinology. 2013;66:31-6. 419. Becker C, Reiter M, Pfaffl MW, Meyer HHD, Bauer J, Meyer KHD. Expression of immune relevant genes in pigs under the influence of low doses of deoxynivalenol (DON). Mycotoxin Research. 2011;27(4):287. 420. Ibáñez-Vea M, González-Peñas E, Lizarraga E, López de Cerain A. Co-occurrence of mycotoxins in Spanish barley: A statistical overview. Food Control. 2012;28(2):295- 8. 421. Ibáñez-Vea M, González-Peñas E, Lizarraga E, López de Cerain A. Co-occurrence of aflatoxins, ochratoxin A and zearalenone in barley from a northern region of Spain. Food Chemistry. 2012;132(1):35-42. 422. Ibáñez-Vea M, Lizarraga E, González-Peñas E, López de Cerain A. Co-occurrence of type-A and type-B trichothecenes in barley from a northern region of Spain. Food Control. 2012;25(1):81-8. 423. Hooper DG, Bolton VE, Guilford FT, Straus DC. Mycotoxin detection in human samples from patients exposed to environmental molds. International journal of molecular sciences. 2009;10(4):1465-75. 424. Brewer JH, Thrasher JD, Straus DC, Madison RA, Hooper D. Detection of mycotoxins in patients with chronic fatigue syndrome. Toxins. 2013;5(4):605-17. 425. Ringot D, Chango A, Schneider YJ, Larondelle Y. Toxicokinetics and toxicodynamics of ochratoxin A, an update. Chemico-biological interactions. 2006;159(1):18-46. 426. Mersch-Sundermann V, Knasmuller S, Wu XJ, Darroudi F, Kassie F. Use of a human- derived liver cell line for the detection of cytoprotective, antigenotoxic and cogenotoxic agents. Toxicology. 2004;198(1-3):329-40. 427. Zhang X, Jiang L, Geng C, Cao J, Zhong L. The role of oxidative stress in deoxynivalenol-induced DNA damage in HepG2 cells. Toxicon : official journal of the International Society on Toxinology. 2009;54(4):513-8.

195

REFERENCES

428. Dohnal V, Wu Q, Kuca K. Metabolism of aflatoxins: key enzymes and interindividual as well as interspecies differences. Archives of toxicology. 2014;88(9):1635-44. 429. McLean M, Dutton MF. Cellular interactions and metabolism of aflatoxin: an update. Pharmacology & therapeutics. 1995;65(2):163-92. 430. He CH, Fan YH, Wang Y, Huang CY, Wang XC, Zhang HB. The individual and combined effects of deoxynivalenol and aflatoxin B(1)on primary hepatocytes of Cyprinus carpio. International journal of molecular sciences. 2010;11(10):3760-8. 431. Corcuera LA, Arbillaga L, Vettorazzi A, Azqueta A, Lopez de Cerain A. Ochratoxin A reduces aflatoxin B1 induced DNA damage detected by the comet assay in Hep G2 cells. Food and chemical toxicology : an international journal published for the British Industrial Biological Research Association. 2011;49(11):2883-9. 432. Li Y, Zhang B, He X, Cheng W-H, Xu W, Luo Y, et al. Analysis of Individual and Combined Effects of Ochratoxin A and Zearalenone on HepG2 and KK-1 Cells with Mathematical Models. Toxins. 2014;6(4):1177-92.

196

ANNEX - SUPPLEMENTARY MATERIAL

ANNEX -. SUPPLEMENTARY MATERIAL

APPENDIX I

Table AI.1 – Information regarding the formula, molecular weight, purity and supplier of all reagents and specific materials used in this thesis. Molecular Molecular weight Purity (%) Supplier formula (g/mol) Gibco/Life 0.25% Trypsin-EDTA solution technolgy 1,3-cyclohexanedione (1,3- C6H8O2 112.13 97 Sigma-Aldrich CHD)

2-thiobarbituric acid (TBA) C4H4N2O2S 114.15 >98 Panreac 2,6-Di-tert-butyl-4-methyl- C15H24O 220.35 99 Sigma-Aldrich phenol (BHT) 75-cm2 culture flasks Corning

Acetic acid C2H4O2 60.05 >98 Merck HPLC Diverse Acetonitrile (ACN) C2H3N 41.05 grade suppliers LC-MS ACN C2H3N 41.05 Riedel-de Häen grade Ammonium acetate C2H7NO2 77.08 98 Panreac

Ammonium carbonate (NH4)2CO3 96.09 Panreac Anhydrous Magnesium MgSO4 120.37 >98 Sigma-Aldrich sulphate

Anhydrous sodium sulphate Na2SO4 142.04 >99 Merck

Bile extract (B8631) Sigma-Aldrich Bovine Serum Albumin (BSA) ~66000 >98 Sigma-Aldrich

C18 column (100 mg) Agilent

Calcium chloride CaCl2 110.98 97 Panreac Chymotrypsin (50 U/mg, Sigma-Aldrich C4129)

Cyclohexane C6H12 84.16 >99 Carl Roth Gibco/Life DEMEM mediumd technolgy Diverse Dichloromethane CH2Cl2 84.93 suppliers

Dimethyl sulfoxide (DMSO) (CH3)2SO 78.13 99.5 Sigma-Aldrich Disodium hydrogen Na2HPO4 141.96 >99 Merck phosphate

198

ANNEX - SUPPLEMENTARY MATERIAL

Table AI.1 – (continued)

Molecular Molecular Purity Supplier formula weight (g/mol) (%) Ethanol C2H6O 46.06 Diverse suppliers Ethyl acetate C4H8O2 88.11 >99 Merck

Ethylenediamine tetraacetic acid C10H16N2O8 292.2 >99 J.T. Baker (EDTA) Foetal bovine serum (FBS) Gibco/Life technolgy Formic acid CH2O2 46.03 >98 Merck

Guanidine-HCl CH6ClN3 95.53 >98 Sigma-Aldrich

GlutaMAXTM 100x Gibco/Life technolgy HBSS mediumb Gibco

a HEPES C8H18N2O4S 238.3 99.5 Sigma-Aldrich

Heptane C7H16 100.2 >99 Carl Roth

Hydrochloride acid HCl 36.46 37.2 Diverse suppliers Isopropanol C3H8O 60.1 >99 Carl Roth

Lipase from porcine pancreas (1000 Sigma-Aldrich U/mg, L3126) Magnesium chloride hexahydrate MgCl2.(H2O)6 203.3 99 Panreac

MEM NEAA 100xe Gibco/Life technolgy Methanol (MeOH) CH4O 32.04 HPLC Diverse grade suppliers MeOH CH4O 32.04 LC-MS Riedel-de grade Häen Pancreatin from porcine pancreas Sigma-Aldrich (P1750) Pefabloc ®SC 100 mg (Protease Sigma-Aldrich inhibitor) Penicillin/Streptomycin 100x solution Gibco/Life (10.000 Units ml-1/10.000 ug ml-1) technolgy Pepsin (2500 U/mg, P7012) Sigma-Aldrich Porcine α-amylase Sigma-Aldrich

Potassium carbonate K2CO3 138.21 Sigma-Aldrich Potassium chloride KCl 74.55 >99 Merck

Potassium dihydrogen phosphate KH2PO4 136.09 Merck Rabbit gastric extract (25 U/mg gastric Lypolytech lipase; 800 U/mg pepsin) RPMI 1640 mediumc Gibco/Life technolgy

199

ANNEX -. SUPPLEMENTARY MATERIAL

Table AI.1 – (continued)

Molecular Molecular Purity Supplier formula weight (g/mol) (%) Sodium dihydrogen phosphate NaH2PO4 119.98 >98 Merck

Sodium hydroxide NaOH 39.99 >97 Merck

Sterilised 96-well plates TPP

Sulphuric acid H2SO4 98.08 96 Diverse suppliers Syringe filters (PES, 0.2 um x 33 mm) TPP

Thiazolyl Blue Tetrazolium Bromide C18H16BrN5S 414.32 98 Sigma- (MTT) Aldrich Transwell permeable polycarbonate Corning supports and inserts (24 mm diameter, 0.4 um pore size) Trichloroacetic acid (TCA) C2HCl3O2 163.38 Panreac Trypsin (13000 U/mg, T0303) Sigma- Aldrich Tryptan blue solution 0.4% C34H24N6Na4O14S4 960.81 Sigma- Aldrich Water H2O 18.01 LC-MS J. T. Baker grade aHEPES: 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid; bHBSS - Hank's Balanced Salt Solution; cRPMI - Roswell Park Memorial Institute; dDMEM - Dulbecco's Modified Eagle Medium; eMEM NEAA – Minimum Essential Medium Non-Essential Amino Acids

200

ANNEX - SUPPLEMENTARY MATERIAL

201

ANNEX -. SUPPLEMENTARY MATERIAL

202

ANNEX - SUPPLEMENTARY MATERIAL

203

ANNEX -. SUPPLEMENTARY MATERIAL

204

ANNEX - SUPPLEMENTARY MATERIAL

205

CHAPTER VI – SUPPLEMENTARY MATERIAL

Texture profile analysis (TPA)

Chicken breast muscle texture analysis was performed using a texture analyzer (Model TA- XT2iHR; Stable Micro System, Ltd, Survey, UK) with a 5 kg load cell to evaluate the stress applied to the samples during perforation. Calibrations were performed with a 2 kg load cell. Exponent software supplied with the instrument was used. Meat cubs (approximately 1 cm3) were subjected to a 2 mm penetration depth through a two-cycle sequence using a 4 mm probe (P/4) with pre-test speed of 3.00 mm s-1, a test (compression) speed of 0.5 mm s-1; and a post-test speed of 3.00 m s-1. The texture parameter assessed were hardness (N) = maximum force required to compress the sample (peak force during the first compression cycle); springiness (m) = height that the samples recovers during the time that elapses between the end of the first compression and the start of the second; cohesiveness (dimensionless) = extent to which the sample could be deformed before rupture (A1/A2, A1 being the total energy required for the first compression and A2 the total energy required for the second compression); and chewiness (J) = the work needed to chew a solid sample to a steady state of swallowing (hardness x cohesiveness x springiness). The results and optimized conditions are showed in Figure AI.1.

206

CHAPTER VI – SUPPLEMENTARY MATERIAL

Figure AI.1 – Texture profile analysis (TPA): A – typical graph obtained showing the first and second bite and the parameters measured; B – Representation of principal component analysis (PCA) and hierarchical cluster analysis (dendrogram) performed to infer which condition better simulates the control condition (F_Ce and F_CI for oven, and M_CE and M_CI for microwaving).

207

CHAPTER VI – SUPPLEMENTARY MATERIAL

APPENDIX II

Figure AII.1 – ANCOVA analysis. Means comparison of ACDs percentage after microwave and oven cooking, regardless the other variables studied with a confidence interval of 95%.

208