Ochratoxin A determination in swine muscle and liver from French conventional or organic farming production systems

Accepted Manuscript Ochratoxin A determination in swine muscle and liver from French conventional or organic farming production systems

Vincent Hort, Marina Nicolas, Brice Minvielle, Corentin Maleix, Caroline Desbourdes, Frédéric Hommet, Sylviane Dragacci, Gaud Dervilly-Pinel, Erwan Engel, Thierry Guérin PII: DOI: Reference:

S1570-0232(18)30122-3 doi:10.1016/j.jchromb.2018.05.040 CHROMB 21204

To appear in: Received date: Revised date: Accepted date:

19 January 2018 19 April 2018 27 May 2018

Please cite this article as: Vincent Hort, Marina Nicolas, Brice Minvielle, Corentin Maleix, Caroline Desbourdes, Frédéric Hommet, Sylviane Dragacci, Gaud Dervilly-Pinel, Erwan Engel, Thierry Guérin , Ochratoxin A determination in swine muscle and liver from French conventional or organic farming production systems. (2017), doi:10.1016/ j.jchromb.2018.05.040

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT Ochratoxin A determination in swine muscle and liver from French conventional or organic farming production systems

Vincent Horta, Marina Nicolasa, Brice Minvielleb, Corentin Maleixa, Caroline Desbourdesa,

b c

RI

Université Paris-Est, Anses, Laboratory for Food Safety, F-94701 Maisons-Alfort, France. IFIP-institut du porc, La Motte au Vicomte, F-35650 Le Rheu, France

SC

a

PT

Frédéric Hommeta, Sylviane Dragaccia, Gaud Dervilly-Pinelc, Erwan Engeld, Thierry Guérina*

LUNAM Université, ONIRIS, Laboratoire d’Etude des Résidus et Contaminants dans les

d

NU

Aliments (LABERCA), Nantes, F-44307, France

INRA, UR370 QuaPA, Microcontaminants, Aroma & Separation Science group (MASS), F-

PT E

D

MA

63123 Saint-Genès-Champanelle, France

CE

Declarations of interest: none

AC

*Corresponding author. Tel.: +33 149 772 711; E-mail address: [email protected]

1

ACCEPTED MANUSCRIPT Abstract Consumers generally considered organic products to be healthier and safer but data regarding the contamination of organic products are scarce. This study evaluated the impact of the farming system on the levels of ochratoxin A (OTA) in the tissues of French pigs (muscle and liver) reared following three different types of production (organic, Label Rouge and

PT

conventional). Because OTA is present at trace levels in animal products, a sensitive ultra-

RI

high performance liquid chromatography–tandem mass spectrometry method using stable isotope dilution assay was developed and validated. OTA was detected or quantified (LOQ of

SC

0.10 µg kg-1) in 67% (n = 47) of the 70 pig liver samples analysed, with concentrations

NU

ranging from < 0.10 to 3.65 µg kg-1. The maximum concentration was found in a sample from organic production but there were no significant differences in the content of OTA between

MA

farming systems. OTA was above the LOQ in four out of 25 samples of the pork muscles. A good agreement was found between OTA levels in muscle and liver (liver concentration = 2.9

PT E

D

x OTA muscle concentration, r = 0.981).

Keywords: Ochratoxin A, pig tissues, conventional meat, organic meat, SIDA–UHPLC–

AC

CE

MS/MS

2

ACCEPTED MANUSCRIPT 1. Introduction Ochratoxin A (OTA) is a mycotoxin produced as a secondary toxic metabolite by various fungi of the genera Apergillus and Penicillium [1]. This mycotoxin has been considered as a possible cause of the human disease known as Balkan Endemic Nephropathy (BEN). In 1993, the International Agency for Research on Cancer (IARC) classified it as possibly carcinogenic

PT

to humans (group 2B) [2]. OTA is most likely produced during storage (post-harvest

RI

formation) under conditions that favour the growth of mould [3]. Cereals and cereal products,

SC

followed by coffee, wine and beer, are assumed to be the major dietary source of exposure. Nevertheless, OTA may also enter the food chain via animal products. In this last case, the

NU

carry-over from naturally contaminated feed should be considered the contamination route of major concern [4]. Among farmed animals, pigs are known to be particularly sensitive to

MA

OTA [3]. Progressive nephropathy is observed in pigs at dietary concentrations of 1 mg kg-1. After absorption through the gastrointestinal tract, the highest concentrations of OTA in pigs

D

are found in the blood, followed by the kidneys, liver, muscle and fat [5]. On the basis of the

PT E

lowest observed adverse effect level (LOAEL) of 8 µg kg-1 body weight (b.w.) per day for early markers of renal toxicity in pigs (the most sensitive animal species), a Tolerable Weekly

CE

Intake (TWI) of 120 ng kg-1 b.w. was derived for OTA by EFSA [6].

AC

OTA occurrence in pig tissues has been widely studied in Europe [5,7–11], but there are only few data regarding France [12,13]. The European Commission established maximum admissible levels for mycotoxins in several foodstuffs (cereals, cereal products, dried vine fruits, coffee, wine, and spices) but not in animal products [14]. However, some countries have enforced or recommended maximum levels of OTA concentration, such as Denmark (pig kidney 10 µg kg-1), Estonia (pig liver 10 µg kg-1), Romania (meat 5 µg kg-1), Slovakia (meat and milk 5 µg kg-1) and Italy (pig meat and derived products 1 µg kg-1) [4].

3

ACCEPTED MANUSCRIPT According to the scientific opinion on OTA contamination of the European Food Safety Authority [3], more extensive occurrence data on ochratoxin A in animal tissues and products thereof, covering all Member States, are required to assess the significance of residue levels. Moreover, an increasing proportion of the European population shows a preference for organically produced agricultural products due to the alleged absence of chemical

PT

contaminants within this mode of production. However, regarding mycotoxin content, it is

RI

feared that products from organic agriculture could be more affected than those from

SC

conventional agriculture, due to the lack of fungicide applications that may thus facilitate fungal infections [15,16]. But there is still not much information in the literature about the

NU

impact of the type of agricultural system on the mycotoxin content in animal products [13,17– 19]. Concerning OTA, no significant difference was observed between conventional and

MA

organic production [13,17,18], excepted in the study of Pozzo et al. [19]. Because OTA is present at trace levels in animal products, sensitive and accurate analytical

D

methods are required. In the last few years, the new generation of tandem MS/MS instruments

PT E

have enabled detection and quantification limits comparable to those obtained with fluorescence detectors [7,10,20–26], stimulating the development of ultra-high performance

CE

liquid chromatography–tandem mass spectrometry (UHPLC–MS/MS) methods [27–33]. The implementation of stable isotope dilution assays (SIDA) enabled an optimal compensation for

AC

losses of OTA in all analytical steps. Although SIDA and UHPLC–MS/MS have been used in combination for OTA determination in pig meat [34], to the best of our knowledge, it has for the first time been applied in 2018 for the analysis of edible pork offal by Cao et al. [32]. The aim of this paper was to specifically assess and compare the OTA content between conventional and organic pig edible tissues [13], implementing a sensitive SIDA–UHPLC– MS/MS method. The correlation between the OTA content in paired liver and meat samples was also investigated. 4

ACCEPTED MANUSCRIPT

2. Material and methods 2.1. Selection and preparation of samples The French main farming practices during the growing-finishing stage were taken into account: organic pig production (straw bedding and outdoor access), Label Rouge (straw

PT

bedding and outdoor access), and conventional production (indoor, on slatted floor). Label

RI

Rouge is a standard farming production system, but intermediate between organic and

SC

standard production in terms of practices and zootechnical criteria such as outdoor access, age and weight at slaughter [35]. Sampling was conducted in 6 slaughterhouses over a 5-month

NU

period, which began in March 2014. One to five visits per slaughterhouse were organised, with a total of 16 days of collection, and from 2 to 13 pig herds sampled per slaughter day.

MA

For each selected herd, three carcasses were randomly chosen from one batch of pigs slaughtered on a single day after post mortem veterinary validation. The average age of the

D

pigs at slaughter was around 26 weeks (from 22 to 28 weeks, according to production system,

PT E

and optimal carcass grading for farmers) while the carcass weight averaged around 95 kg. A pooling strategy (1 sample = 1 pool of 3 individuals from the same farm) was operated in

CE

order to assess the average level of the group-housed growing-finishing pigs, thus overcoming any possible individual variability.

AC

On each animal, two types of edible tissues representative of different type of chemical contamination were sampled [13]: the tenderloin (psoas major muscle) was removed from the carcass (average weight of 544 g), and a sample of liver tissue was taken from the half distal part of the right lateral lobe (average weight of 174 g). All samples were individually packed in plastic bags and refrigerated at 2°C. On the same day, all samples were homogenised, ground using a meat grinder and pooled by farm, then frozen at -20°C in polypropylene jars. Before use, all utensils and vessels were washed with tap water, immerged half-an-hour in

5

ACCEPTED MANUSCRIPT water added with RBS detergent (2% solution v/v) and successively rinsed with tap water, purified water and ultrapure water. Sampling frames were set up such that the identities of the slaughterhouses and farms of origin were anonymous. In all, pooled samples of muscle and liver pairs were collected from 28, 12 and 30 organic, Label Rouge and conventional farms, respectively, from different regions of France

RI

PT

representative of the French pig production [13].

SC

2.2. Chemicals and reagents

All solutions were prepared with analytical reagent-grade chemicals and ultrapure water (18.2

NU

MΩ cm) obtained by purifying distilled water with the Milli-Q system associated with an Elix 5 pre-system (Millipore S.A., St Quentin-en-Yvelines, France). Ochratoxin A standard (OTA)

MA

was purchased from Sigma (St. Louis, MO, USA). Stock solution of OTA (100 µg mL-1) was made by dissolving pure standard in methanol and the concentration was determined by

D

absorption spectrometry (Ɛ = 6330 L.mol-1.cm-1). Two working standard solutions (100 and 5

PT E

ng mL-1) prepared in methanol were used to spike blank matrix (a muscle or liver matrix that does not contain detectable OTA) and for the preparation of calibration standards.

CE

Internal standard solutions of uniformly labelled U-[13C20]-OTA was purchased from Biopure (Tulln, Austria) (10 µg mL-1). A stock standard solution (500 ng mL-1) was prepared in

AC

methanol. A working standards solution of U-[13C20]-OTA (56 ng mL-1) was added to all samples, blanks and calibration standards. Calibration standards were prepared by diluting OTA and U-[13C20]-OTA working standard solutions with water/methanol/formic acid 75/22/3 (v/v/v). The following OTA concentrations were obtained: 0.000, 0.125, 0.500, 1.25 and 2.50 ng mL-1. The U-[13C20]-OTA concentration was the same for all the levels (1.4 ng mL-1)

6

ACCEPTED MANUSCRIPT HPLC grade acetonitrile, methanol, ethyl acetate and acetic acid were purchased from Fisher Scientific (Loughborough, UK). 85% orthophosphoric acid (Fisher Scientific), sodium hydrogen carbonate (Merck KGaA, Darmstadt, Germany), formic acid (Fisher Scientific) and ammonium formate (Alfa Aesar, Karlsruhe, Germany) were of analytical grade. Besides calibration standard preparation, water/methanol/formic acid 75/22/3 (v/v/v) was

PT

used to adjust the final volume before UHPLC–MS/MS injection. An orthophosphoric acid

RI

1M solution was prepared by diluting 34 mL of 85% orthophosphoric acid in 500 mL of

SC

water. A sodium hydrogen carbonate 0.5 M solution was prepared by dissolving 42 g sodium hydrogen carbonate in 1 L of water. Ultra-pure grade carrier argon (Ar, 99.9999% pure) and

NU

nitrogen (N2, 99.999% pure) were purchased from Linde Gas (Montereau-Fault-Yonne,

MA

France).

2.3. Extraction

D

The extraction procedure followed was a modification of the method developed by Matrella et

PT E

al. [8]: after the double extraction with acidic ethyl acetate, and a liquid-liquid clean-up with 0.5 M sodium hydrogen carbonate, we introduced an immunoaffinity clean-up and stable

CE

isotope dilution assays (SIDA). Immunoaffinity limits matrix interferences, which is critical in complex matrices, and SIDAs enable an optimal compensation for losses of ochratoxin A

AC

in all analytical steps. A test portion of 2.50 ± 0.01 g of muscle or liver sample was acidified with 750 µL of orthophosphoric acid 1M. The tube was then mixed using a vortex mixer. After 5 minutes, 25 µL of U-[13C20]-OTA working standard solution were added to the test portion. The tube was then mixed again using a vortex mixer. 10 mL ethyl acetate were added and the samples were homogenized at 10000 ± 500 rpm for 2 min using a Polytron® (Kinematica AG, Luzern, Switzerland). After centrifugation at 9000 g for 10 min at 3°C, the

7

ACCEPTED MANUSCRIPT supernatant was placed in a funnel. The extraction of OTA from the test portion with 10 mL ethyl acetate and the centrifugation step were repeated once.

2.4. Clean-up Two clean-up steps were implemented (for both muscle and liver samples). First a liquid-

PT

liquid purification was carried out with 10 mL of 0.5 mol/L sodium hydrogen carbonate added

RI

to the 20 mL of extract. The funnel was shaken vigorously three times, for 15 s each time, and

SC

the tap was opened to release excess vapour pressure. The mixture was left to stand for 5 min and the aqueous phase was then placed in a 50 mL tube. The liquid-liquid purification was

NU

repeated a second time and followed by an immunoaffinity clean-up. The aqueous phase was slowly (drop by drop) passed through an Ochraprep® immunoaffinity column (R-Biopharm

MA

Rhône Ltd, Glasgow, Scotland), placed in a vacuum manifold and washed with 20 mL of water. The column was then dried by pushing air through it with a syringe. Ochratoxin A was

D

finally eluted twice with 1.5 mL methanol/acetic acid 98/2 (v/v).

PT E

The eluate was evaporated at 50 ± 5°C under a gentle stream of nitrogen to 0.1 mL and adjusted to 1 mL with water/methanol/formic acid 75/22/3 (v/v/v). The final extract was

CE

mixed using the vortex mixer, filtered (PET, 20 µm) and analysed by UHPLC–MS/MS.

AC

2.5. UHPLC conditions

The UHPLC system was an Accela 1250 (Thermo Fisher Scientific, San Jose, CA, USA). To optimize the chromatographic separation, the following three columns were tested: Hypersil GOLD® C18 50 mm x 2.1 mm, 1.9 µm (Thermo Scientific, Waltham, MA, USA ), Kinetex® C18, 50 mm x 2.1 mm, 1.7 µm and Kinetex® C18, 50 mm x 2.1 mm, 2.6 µm (Phenomenex, Torrance, CA, USA). The final chromatographic separation of mycotoxins was performed using a column with particles composed of a solid core and a porous shell. A Kinetex® C18

8

ACCEPTED MANUSCRIPT column (100 Å, 2.6 µm particle size, 50 x 2.1 mm) equipped with a Kinetex C18 security guard cartridge was used. Eluent A was composed of water/formic acid 99.9/0.1 (v/v) and eluent B of methanol/water/formic acid 94.9/5/0.1 (v/v/v). Both eluents contained 0.5 mmol/L ammonium formate. The gradient was programmed as follows: 25% B (initial), 25–100% B (6.4 min), 100% B (hold 1.1 min), 100–25% B (0.5 min), 25% B (hold 1 min). The column

PT

effluent was transferred via a divert valve (Rheodyne, USA) either to the mass spectrometer

RI

(between 1.5 and 5.5 min) or to waste. The total flow rate was 0.5 ml/min while the injection volume was 10 µL. The column temperature was maintained at 30°C. A chromatogram of a

NU

SC

blank matrix spiked at the quantification limit (0.10 µg kg-1) is shown in Fig. 1.

2.6. MS/MS conditions

MA

The detection system was performed with a TSQ Vantage triple quadrupole mass spectrometer (Thermo Fisher Scientific) equipped with an ESI source (HESI-II probe). The

D

mass spectrometer was operated in positive ESI mode. The source temperature was set at

PT E

500°C and capillary temperature at 350°C. The spray voltage was 4000 V. Air was used as nebulising gas with a sheath gas pressure of 50 (arbitrary unit) and an auxiliary gas pressure

CE

of 18 (arbitrary unit). The collision gas was argon with a gas pressure of 1.5. S-lens was set to 80 V. The mass spectrometer was operated in selected reaction monitoring (SRM) mode. For

AC

OTA and U-[13C20]-OTA, one transition was used for quantification (Q) and another as qualifier transition (q). The following transitions were monitored: OTA (Q): m/z 404.1 → 239.0 (collision energy, CE = 23 V), OTA (q): m/z 404.1 → 102.1 (CE = 24 V); U-[13C20]OTA (Q): 424.1 → 250.0 (CE = 64 V); U-[13C20]-OTA (q): m/z 424.1 → 231.9 (CE = 33 V). A mass resolution of 0.7 Da full width at half maximum (FWHD) was set on the first (Q1) and the third (Q3) quadrupoles. Instrument control and data were handled by a computer

9

ACCEPTED MANUSCRIPT equipped with TSQ Tune Master version 2.3.0, Xcalibur version 2.1.0 and TraceFinder version 1.0.1 (Thermo Fisher Scientific).

2.7. Method validation methodology The method was validated according to the accuracy profile procedure, based on tolerance

PT

intervals to select the best calibration function and to determine the validated concentration

RI

ranges [36,37]. This procedure gives a precise estimation of the accuracy of the analytical

SC

method and provides a realistic limit of quantification. The accuracy profile summarises every validation element on a single plot, giving a graphical representation of the error risk for each

NU

concentration in the validated range and of the method’s performance. The accuracy profile expresses the total error including systematic error (trueness) and random error (repeatability

MA

and intermediate precision) for each concentration level. The validity domain is defined between the lowest and the highest tested concentrations with tolerance limits (β-expectation

D

limits) between the acceptance limits (λ). Tolerance limits were calculated at each

PT E

concentration level and take into account the reference value, the bias and the intermediate precision. A multi-matrixes design of experiment was built according to the NF V03-110

CE

standard [37] with 3 series of meat and 3 series of liver. The limit of quantification (LOQ)

AC

was defined as the lowest concentration level validated for both matrices.

2.8. Quality control

To ensure reliable results, samples were analysed in batches including several internal quality controls (IQCs). When acceptance criteria were not met, results were discarded and samples were re-analysed. A blank matrix of liver or muscle was analysed in the same conditions as for all samples, to check interferences. Concentration had to be below the detection limit. A blank matrix spiked at the quantification limit (0.10 µg kg-1) was prepared and analysed in the

10

ACCEPTED MANUSCRIPT same conditions as for all the samples. Recoveries were to be between 30 and 140% without isotopic dilution (external calibration) and between 70 and 130% by using isotopic dilution. OTA was quantified using bracketing calibration curves composed of 5 levels at 0.000, 0.125, 0.500, 1.25 and 2.50 ng mL-1. For each calibration level, the concentration of U-[13C20]-OTA was 1.4 ng mL-1. The determination coefficient (r²) of the calibration curve had to be ≥ 0.98

PT

and slope variation between two sets of bracketing calibration curves had to be below 15%.

RI

Variation of the retention time of OTA in samples had to be below 5% in comparison to

SC

standard retention time. Before use, each batch of immunoaffinity columns was checked. The

NU

column capacity had to be higher than 100 ng and the column recovery higher than 85%.

2.9. Statistical data analysis

MA

In order to deal with the left-censored (LC) data, the substitution method described by WHO and EFSA [38,39] was implemented. As recommended, data between the limit of detection

D

(LOD) and the LOQ were quantified for the purpose of dietary exposure assessments. Lower

PT E

bound (LB) and upper bound (UB) scenarios were calculated. In a lower bound scenario, nondetected and non-quantified results were replaced by zero. In an upper bound scenario, non-

CE

detected results were replaced by the LOD and results higher than the LOD were used. When the difference between the lower and upper bound values is negligible, the upper bound value

AC

was reported. Statistical data analyses were carried out in order to compare different types of farming. After implementation of a statistical normality test (Shapiro-Wilk), a median comparison test was therefore performed. The non-parametric Mann Whitney test was implemented for this purpose with a significance level α = 5%. Correlation between OTA levels in muscle and liver was also studied using a Pearson statistical test with a significance level α = 5%. Statistical data analysis was performed using Stagraphics Centurion XVI version 16.1.17 (StatPoint Technologies, Inc., Warrenton, VA, USA).

11

ACCEPTED MANUSCRIPT

3. Results and discussion 3.1. Optimization of analytical conditions Animal products are particularly challenging matrices, posing difficult problems in analysis. In this study, problems are compounded by the strong binding between proteins and OTA in

PT

addition to the need to remove interferences that could be co-extracted [29]. To optimize the

RI

chromatographic separation, three columns were tested: Hypersil GOLD® C18 50 mm x 2.1

SC

mm, 1.9 µm, Kinetex® C18, 50 mm x 2.1 mm, 1.7 µm and Kinetex® C18, 50 mm x 2.1 mm, 2.6 µm. The peak widths at half of their height and the peak intensities achieved were similar

NU

whatever the column (ranging from 5100 to 5340 cps) in the best conditions. However, the Kinetex® C18 column, 2.6 µm particle size was chosen because this column is composed of

MA

superficially porous particles allowing faster analysis and greater efficiency with less backpressure than fully porous particle columns. Performances similar to those provided by

3.2. Method validation

PT E

D

columns with fully porous particles, designed for UHPLC (sub 2 µm), were obtained.

CE

To validate the method, an accuracy profile was built following the NF V03-110 standard: 2010 [36,37]. This includes a range of 3 concentration levels (0.10, 0.30 and 1.0 µg kg-1) and

AC

6 series repeated on 6 different days, over 6 weeks. To estimate the first level (LOQ level), some samples were spiked at levels lower or equal to 0.10 µg kg-1. For each series, 2 replicates of each of the 3 concentration levels were analysed in bracketing with five calibration standards to establish the response function. The probability β was set to 80%, meaning that the risk of results falling outside β-expectation tolerance intervals was below 20% on average. The acceptance limits (λ) were set at ± 25% according to the expected performance of the method, initially defined. The accuracy profile and the performance

12

ACCEPTED MANUSCRIPT criteria obtained for OTA are presented in Fig. 2 and Table 1. For all the concentration levels, β-expectation tolerance intervals were within the acceptability limits. The recovery ranged from 95% to 105%. The estimated repeatability coefficient of variation, CVr, varied from 3.8% to 8.5% on the validity domain. The intermediate precision coefficient of variation, CVIP, varied from 8.0% to 11%. The expanded uncertainty of 23% was calculated using data

PT

from the accuracy profile.

RI

The method developed features very high sensitivity with a LOQ corresponding to the first level of concentration validated (0.10 µg kg-1).

SC

The validity domain of the analytical method ranged from 0.10 to 1.0 µg kg-1.

NU

In the past, fluorescence detection was mainly applied for OTA determination in animal tissues. According to the ten methods listed by Duarte et al. in their review [40] and the most

MA

recent studies [22–26], only 4 fluorimetric methods allow obtaining lower LOQ than our method [7,10,20,41]. However, the recoveries and precision of our method are significantly

D

improved. The LOQ observed in this study (0.10 µg kg-1) is similar or lower than those

PT E

estimated by others recent studies using LC-MS/MS methods (range 0.10 – 1.0 µg kg-1) [28–

CE

33].

3.3. Occurrence of OTA in pig liver samples

AC

Among the 70 pig livers analysed, OTA was quantified in 36% (n = 25) and detected in 31% (n = 22) of the samples (Table 2). Three samples had OTA content above 1.0 µg kg-1 (1.2, 1.5 and 3.7 µg kg-1) which are the recommended maximum OTA levels in pig and pig-derived products in Italy. However, those values are far lower than the maximum OTA level observed by Armorini et al. [23] in a salami sample (103.7 µg/kg). The difference between LB and UB scenarios is negligible (0.03 µg kg-1). The UB mean concentration of 0.18 µg kg-1 concurs with those previously reported (range 0.16–0.20 µg kg-1) [5,42,43]. However, higher results

13

ACCEPTED MANUSCRIPT were also reported e.g. a mean LB of 0.63 µg kg-1 from 90 liver samples with a maximum OTA level of 14.5 µg kg-1 [9]. In the study of Hakim et al. [44], OTA was quantified in all the liver samples with a mean value of 3.78 µg kg-1 sampled from Egyptians markets. Our results are in the same range than those obtained with pig meat in some recent studies [26,29]. However, it should be noted that factors such as climate conditions during harvest,

PT

practices for grain drying and storage, and feed composition, etc. influence the OTA levels

SC

RI

found in pig tissues [9].

3.4. Comparison of OTA contamination in pig liver between different types of farming

NU

Considering the 70 pig liver samples analysed belonging to 3 types of farming: conventional, Label Rouge and organic productions, the maximum concentration (3.65 µg kg-1) was found

MA

in a sample from organic production. In Label Rouge and conventional productions, the highest concentrations were 1.49 and 0.29 µg kg-1, respectively (Table 2). As the differences

PT E

different types of farming [39]

D

between LB and UB scenarios were negligible, only UB scenarios were used to compare the

Based on UB mean concentration, samples from animals raised according to organic

CE

specifications were three times more contaminated by OTA than those from conventional animals (Fig. 3). Label Rouge UB mean concentration (n = 12) is between those obtained for

AC

conventional and organic samples. A possible explanation could be the fact that Label Rouge is an intermediate farming production system in term of practices and zootechnical criteria. Regarding UB median concentration, conventional and organic are very similar and lower than Label Rouge. The difference between mean and median could be explained by a small number of more heavily contaminated samples.

14

ACCEPTED MANUSCRIPT A non-parametric Mann-Whitney statistical test was carried out with the UB scenarios in order to determine whether there are significant differences between the farming production systems. Conventional and organic productions were not significantly different (P-value = 0.708). Label Rouge production was not compared statistically with the other two types of farming

PT

because the number of results is below 25 (n = 12). However, because Label Rouge is an

RI

intermediate type of farming, results were merged, first to the organic samples and compared

SC

to the conventional production (P-value = 0.846) then to the conventional production and compared with the organic production (P-value = 0.654). In both cases P-values were below

NU

the significance level of 5% and no statistical difference was observed. Very few previous studies investigated the impact of the system of animal production in terms of OTA

MA

contamination [13,17–19] and a difference was demonstrated between conventional and

D

organic farming production systems in the study of Pozzo et al. [19].

PT E

3.5. Correlation study between OTA concentration in pig liver and meat Until now, serum and kidney were the most suitable samples for monitoring OTA

CE

contamination in animal products [40]. Indeed, the correlation between serum and kidney and kidneys and muscle in term of OTA content was widely studied [5,9] but no correlation was

AC

estimated between liver and muscle. In this study, the correlation between OTA levels in muscle and liver was estimated from the 25 samples where OTA was quantified in livers. Results showed that for each sample “liver–muscle” pair, the OTA concentration was systematically lower in the muscle than in the liver (Table 3). OTA was above the LOQ in 16% of the pork muscles (n = 4), between LOD and LOQ in 60% (n = 15) and below the LOD in 24% of cases (n = 6). OTA was quantified both in liver and muscle in 4 out of 25samples analysed. The highest OTA concentrations in the muscle and in the liver were from

15

ACCEPTED MANUSCRIPT the same pooled animals (sample No 60). A Pearson statistical test was applied [45] and a very high positive correlation was found (P-value = 1.6 x 10-13; r = 0.981 with a 95 percent confidence interval [0.950–0.993]). The equation linking OTA muscle and liver concentrations is: OTA liver concentration = 2.9 x OTA muscle concentration + 0.0 (Fig. 4).

PT

Previously, [5] reported a strong correlation between OTA concentrations in serum and the

RI

kidney (r = 0.91). Nevertheless, the ratio between OTA content in muscle and in kidney was

SC

found to vary between 10% and 90% [9]. The reason for this variability was ascribed to different factors such as the level of feed contamination, the feeding period, or feeding in

NU

relation to time of slaughtering. Although [5] did not indicate a correlation coefficient between OTA levels in liver and meat, according to their results, the average concentration in

MA

liver was estimated 3.4 times higher than meat, very close to that observed in this study (factor of 2.9).

D

In 2017, a significant difference between liver and meat was reported by Altafini et al. [25]

PT E

with a factor in the range 1.7–2.2. The authors have also compared their results with eight others feeding studies of pigs exposed to OTA. The mean factor from those studies was 2.8,

AC

4. Conclusions

CE

that is in agreement with the factor estimated in our study.

Among the 70 pig liver and 25 muscle samples analysed, OTA was detected or quantified respectively in 47 (67%) and 19 (76%) samples at low concentrations thanks to a sensitive SIDA–UHPLC–MS/MS method, validated following the accuracy profile procedure. Three different types of French farming practices were compared in terms of OTA contamination and no significant differences were found between organic, Label Rouge and conventional productions. However, three samples had OTA content above 1.0 µg kg-1 which is the

16

ACCEPTED MANUSCRIPT recommended maximum OTA levels in pig and pig-derived products in Italy and the maximum concentration (3.65 µg kg-1) was found in a sample from organic production. With kidney, liver is one of the tissues most able to highly accumulate this mycotoxin. In this study, the correlation observed between OTA content in muscle and liver seems to be the strongest observed until now among all edible organs. Consequently, liver analysis should be

PT

an alternative approach to be considered by the risk manager for implementing surveillance

SC

RI

strategy.

Acknowledgments

NU

The research work was funded by the French National Agency for Research (ANR), contract no. ANR-12-ALID-004 “Sécurité sanitaire des viandes issues de l’agriculture biologique”,

AC

CE

PT E

D

MA

SOMEAT (Safety of Organic Meat). http://www.so-meat.fr/

17

ACCEPTED MANUSCRIPT References [1] P. Bayman, J. L. Baker, Ochratoxins: A global perspective, Mycopathologia 162(3) (2006) 215–223. [2] IARC, Some naturally occurring substances: food items and constituents, heterocyclic aromatic amines and mycotoxins, IARC Monographs on the Evaluation of Carcinogenic

PT

Risk of Chemicals to Humans, 56 (1993) 1–521.

RI

[3] European Food Safety Authority (EFSA), Opinion of the Scientific Panel on Contaminants

SC

in the Food Chain on a request from the Commission related to ochratoxin A (OTA) as undesirable substance in animal feed, EFSA Journal 101 (2004) 1–36.

NU

[4] S.C. Duarte, C.M. Lino, A. Pena, Mycotoxin food and feed regulation and the specific case of ochratoxin A: a review of the worldwide status, Food Addit. Contam. 27 (2010)

MA

1440–1450.

[5] V.G. Curtui, M. Gareis, E. Usleber, E. Märtlbauer, Survey of Romanian slaughtered pigs

D

for the occurrence of mycotoxins ochratoxins A and B, and zearalenone, Food Addit.

PT E

Contam. 18(8) (2001) 730–738.

[6] European Food Safety Authority (EFSA), Opinion of scientific panel on contaminants in

CE

the food chain on a request from the Commission related to ochratoxin A in food, EFSA Journal, 365 (2006) 1–56.

AC

[7] K. Jørgensen, A. Petersen, Content of ochratoxin A in paired kidney and meat samples from healthy Danish slaughter pigs, Food Addit. Contam. 19(6) (2002) 562–7. [8] R. Matrella, L. Monaci, M.A. Milillo, F. Palmisano, M. G. Tantillo, Ochratoxin A determination in paired kidneys and muscle samples from swines slaughtered in southern Italy, Food Control 17(2) (2006) 114–117.

18

ACCEPTED MANUSCRIPT [9] D. Milićević, V. Jurić, S. Stefanović, M. Jovanović, S. Janković, Survey of slaughtered pigs for occurrence of ochratoxin A and porcine nephropathy in Serbia, Int. J. Mol. Sci. 9(11) (2008) 2169–2183. [10] E.M. Guillamont, C.M. Lino, M.L. Baeta, A.S. Pena, M.I.N. Silveira, J.M. Vinuesa, A comparative study of extraction apparatus in HPLC analysis of ochratoxin A in muscle,

PT

Anal. Bioanal. Chem. 383(4) (2005) 570–575.

RI

[11] S.C. Duarte, C.M. Lino, A. Pena, A, Ochratoxin A in food and urine: a nationwide

SC

Portuguese two-year study, World Mycotoxin J. 8(1) (2015) 121–132. [12] S. Dragacci, F. Grosso, R. Bire, J.M. Frémy, S. Coulon, A French monitoring

NU

programme for determining ochratoxin A occurrence in pig kidneys, Nat. Toxins 7(4) (1999) 167–173.

MA

[13] G. Dervilly-Pinel, T. Guérin, B. Minvielle, A. Travel, J. Normand, M. Bourin, E. Royer, E. Dubreil, S. Mompelat, F. Hommet, M. Nicolas, V. Hort, C. Inthavong, M. Saint-

D

Hilaire, C. Chafey, J. Parinet, R. Cariou, P. Marchand, B. Le Bizec, E. Verdon, E. Engel,

PT E

Micropollutants and chemical residues in organic and conventional meat, Food Chem. 232 (2017) 218–228.

CE

[14] Commission Regulation (EC) No. 1881/2006 of 19 December 2006 setting maximum levels for certain contaminants in foodstuffs, Official Journal of the European Union

AC

L364 (2006) 5–24.

[15] L. Pussemier, Y. Larondelle, C. Van Peteghem, A. Huyghebaert, Chemical safety of conventionally and organically produced foodstuffs: A tentative comparison under Belgian conditions, Food Control 17(1) (2006) 14–21. [16] J.M. Garcia, P. Teixeira, Organic versus conventional food: A comparison regarding food safety, Food Rev. Int. 33(4) (2016) 424–446.

19

ACCEPTED MANUSCRIPT [17] K. Jørgensen, Survey of pork, poultry, coffee, beer and pulses for ochratoxin A, Food Addit. Contam. 15(5) (1998) 550–554. [18] L. Malmauret, D. Parent-Massin, J.L. Hardy, P. Verger Contaminants in organic and conventional foodstuffs in France, Food Addit. Contam. 19(6) (2002) 524–532. [19] L. Pozzo, L. Cavallarin, D. Nucera, S. Antoniazzi, A. Schiavone, A survey of ochratoxin

PT

A contamination in feeds and sera from organic and standard swine farms in northwest

RI

Italy, J. Sci. Food Agric. 90 (2010) 1467–1472.

SC

[20] T. Toscani, A. Moseriti, A. Dossena, C. Dall’Asta, N. Simoncini, R. Virgili, Determination of ochratoxin A in dry-cured meat products by a HPLC–FLD quantitative

NU

method, J. Chromatogr. B. 855 (2007) 242–248.

[21] C. Dall’Asta, G. Galaverna, T. Bertuzzi, A. Moseriti, A. Pietri, A. Dossena, R.Marchelli,

MA

Occurrence of ochratoxin A in raw ham muscle, salami and dry-cured ham from pigs fed with contaminated diet, Food Chem. 120(4) (2010) 978–983.

D

[22] A. Pietri, A. Gualla, S. Rastelli, T. Bertuzzi,Enzyme-assisted extraction for the HPLC

(2011) 1717–1723.

PT E

determination of ochratoxin A in pork and dry-cured ham, Food Addit. Contam. A 28(12)

CE

[23] S. Armorini, A. Altafini, A. Zaghini, P. Roncada, Ochratoxin A in artisan salami produced in Veneto (Italy), Food Addit. Contam. B 9(1) (2016) 9–14.

AC

[24] C. Luan, L. Wang, F. Chen, S. Wang, L. Zhao, L. Shao, Determination of Ochratoxin A in Pig Muscle Using Dispersive Liquid-liquid Microextraction Combined with HighPerformance Liquid Chromatography, Food Anal. Method. 9(6) (2016) 1490–1494. [25] A. Altafini, S. Armorini, A. Zaghini, L. Sardi, P. Roncada, Tissue distribution of ochratoxin A in pigs after administration of two-levels contaminated diets, World Mycotoxin J. 10 (3) (2017) 263–272.

20

ACCEPTED MANUSCRIPT [26] N.J. Mitchell, C. Chen, J.D. Palumbo, A. Bianchini, J. Cappozzo, J. Stratton, D. Ryu, F. Wu, A risk assessment of dietary Ochratoxin a in the United States, Food Chem. Toxicol. 100 (2017) 265–273. [27] D. Milićević, V. Jurić, S. Stefanović, T, Baltić, S. Janković, Evaluation and validation of two chromatographic methods (HPLC-Fluorescence and LC-MS/MS) for the

PT

determination and confirmation of ochratoxin A in pig tissues, Arch. Environ. Cont. Tox.

RI

58(4), (2010) 1074–1081.

SC

[28] D. Chen, X. Cao, Y. Tao, Q. Wu, Y. Pan, L. Huang, X. Wang, Y. Wang, D. Peng, Z. Liu, Z. Yuan, Development of a sensitive and robust liquid chromatography coupled with

NU

tandem mass spectrometry and a pressurized liquid extraction for the determination of aflatoxins and ochratoxin A in animal derived foods, J. Chromatogr. A 1253 (2012) 110–

MA

119.

[29] S.C. Duarte, C.M. Lino, A. Pena, Novel IAC-LC-ESI-MS2 analytical set-up for

D

ochratoxin A determination in pork, Food Chem. 138 (2013) 1055–1061.

PT E

[30] Y.L. Hou, J. Zhou, Y.P. Li, J.H. Xie, L. Zhou, F.L. Lv, Determination of ochratoxin A in pig kidneys by immunoaffinity cleanup and ultra-high performance liquid

CE

chromatography, J. AOAC Int. 98(6) (2015) 1566–1570. [31] Z. Zhao, N. Liu, L. Yang, Y. Deng, J. Wang, S. Song, S. Lin, A. Wu, Z. Zhou, J. Hou,

AC

Multi-mycotoxin analysis of animal feed and animal-derived food using LC-MS/MS system with timed and highly selective reaction monitoring, Anal. Bioanal. Chem. 407 (2015) 7359–7368. [32] X. Cao, X. Li, J. Li, Y. Niu, L. Shi, Z. Fang, T. Zhang, H. Ding, Quantitative determination of carcinogenic mycotoxins in human and animal biological matrices and animal-derived foods using multi-mycotoxin and analyte-specific high performance

21

ACCEPTED MANUSCRIPT liquid chromatography-tandem mass spectrometric methods, J. Chromatogr. B, 1073 (2018) 191–200. [33] A. Paoloni, M. Solfrizzo, R. Bibi, I. Pecorelli, Development and validation of LCMS/MS method for the determination of Ochratoxin A and its metabolite Ochratoxin α in poultry tissues and eggs, J. Environ. Sci. Health., Part B 53(5) (2018) 327–333.

PT

[34] J. Tam, P. Pantazopoulos, P.M. Scott, J. Moisey, R.W. Dabeka, I. D. K. Richard,

RI

Application of isotope dilution mass spectrometry: determination of ochratoxin A in the Canadian Total Diet Study, Food Addit. Contam. A 28(6) (2011) 754–761.

NU

d’un Label Rouge en « porc » (2009) 1–29.

SC

[35] INAO, Notices techniques définissant les critères minimaux à remplir pour l’obtention

http://www.inao.gouv.fr/content/download/485/3605/version/1/file/NT-porc.pdf.

MA

Accessed 13 Avril 2018

[36] J.M. Mermet, G. Granier, Potential of accuracy profile for method validation in

D

inductively coupled plasma spectrochemistry, Spectrochim. Acta B 76 (2012) 214–220.

PT E

[37] AFNOR, Analyse des produits agricoles et alimentaires - Protocole de caractérisation en vue de la validation d’une méthode d’analyse quantitative par construction du profil

CE

d’exactitude - NF V03-110 (2010). [38] GEMS/Food-EURO, Reliable evaluation of low-level contamination of food. Report on a

AC

workshop in the frame of GEMS/Food-EURO, Kulmbach, Germany (1995). [39] European Food Safety Authority (EFSA), Management of left-censored data in dietary exposure assessment of chemical substances, EFSA Journal 8(3):1557 (2010) 1–96. [40] S.C. Duarte, C.M. Lino, A. Pena, Food safety implications of ochratoxin A in animalderived food products, Vet. J. 192(3) (2012) 286–292.

22

ACCEPTED MANUSCRIPT [41] L. Giacomo, V. Michele, F. Guido, M. Danilo, I. Luigi, M. Valentina, Determination of ochratoxin A in pig tissues using enzymatic digestion coupled with high-performance liquid chromatography with a fluorescence detector, MethodsX 3 (2016) 171–177. [42] Joint FAO/WHO Expert Committee on food additives (JECFA), Safety evaluation of certain mycotoxins in food, Geneva (2001).

PT

[43] European Commission (EC), Assessment of dietary intake by the population in EU

RI

member states, SCOOP task 3.2.7 (2002).

SC

[44] A.S. Hakim, R.M. Alarousy, Incidence of Fungal Infections and Mycotoxicosis in Pork Meat and Pork By- Products in Egyptian Markets Incidence of Fungal Infections and

NU

Mycotoxicosis in Pork Meat and Pork By-Products in Egyptian Markets, Int. J. Biol. Biomol. Agric. Food Biotechnol. Eng. 9(7) (2015) 816–819.

MA

[45] M.T. Puth, M. Neuhäuser, G.D. Ruxton, Effective use of Pearson’s product-moment

AC

CE

PT E

D

correlation coefficient, Anim. Behav. 93 (2014) 183–189.

23

ACCEPTED MANUSCRIPT Figure Captions Fig. 1: Chromatograms obtained for a pig liver sample spiked at the quantification limit (0.10 µg kg-1 OTA): a) OTA quantification transition (Q); b) OTA qualifier transition (q); c) U[13C20]-OTA quantification transition (Q); d) U-[13C20]-OTA qualifier transition (q) Fig. 2: Accuracy profile of OTA (β = 80%; λ = ± 25%)

PT

Fig. 3: Box Plot obtained with results in UB scenario for conventional, Label Rouge and

RI

organic farming production systems (+ = Mean concentration (µg kg-1))

AC

CE

PT E

D

MA

NU

SC

Fig. 4: Correlation between muscle and liver OTA concentrations

24

ACCEPTED MANUSCRIPT Figure 1

PT

a)

RI NU

SC

c)

MA

d)

CE

PT E

D

Time (min)

AC

Intensity (cps)

b)

25

ACCEPTED MANUSCRIPT

AC

CE

PT E

D

MA

NU

SC

RI

PT

Figure 2

26

ACCEPTED MANUSCRIPT Figure 3

PT

Conventional

SC

RI

Label Rouge

0.1

0.0

0.2

NU

Organic

1.5

0.3

3.7

-1

AC

CE

PT E

D

MA

OTA Concentration (µg kg )

27

ACCEPTED MANUSCRIPT

AC

CE

PT E

D

MA

NU

SC

RI

PT

Figure 4

28

ACCEPTED MANUSCRIPT Table 1: OTA performance criteria Trueness LOD (µg kg-1)

LOQ (µg kg-1)

Recovery (%)

0.10 0.30 1.0

0.03

0.10

105 95 102

AC

CE

PT E

D

MA

NU

SC

RI

PT

OTA

Levels (µg kg-1)

Accuracy Precision Intermediate Repeatability precision (% RSD) (% RSD) 8.5 11 6.1 8.0 3.8 8.8

29

ACCEPTED MANUSCRIPT Table 2: Pig liver results

Total number

Number of samples ≥ LOD < LODa and < LOQb 23 22

≥ LOQ

Mean LBc (µg kg-1)

Mean UBd (µg kg-1)

Median UBd (µg kg-1)

Max. Value (µg kg-1)

AC

CE

PT E

D

MA

NU

SC

RI

PT

3.65 All data 70 25 0.15 0.18 0.05 Conventiona 0.29 30 9 11 10 0.06 0.09 0.05 l 1.49 Label Rouge 12 4 3 5 0.18 0.20 0.08 3.65 Organic 28 10 8 10 0.23 0.25 0.05 -1 -1 a: LOD = 0.03 µg kg ; b: LOQ = 0.10 µg kg ; c: LB: lower bound scenario; d: UB: upper bound scenario

30

ACCEPTED MANUSCRIPT Table 3: Results in meat samples from the same pool of animals when a value equal to or higher than the LOQ has been measured in livers Batch number of pooled animals

Type of farming

Liver concentration (µg kg-1)

Muscle concentration (µg kg-1)

AC

CE

PT E

D

MA

NU

SC

RI

PT

34 0.10 0.05a 2 0.11 0.04 a 5 0.13 ≤ 0.03b 26 0.14 0.05 a 24 0.16 ≤ 0.03 b Conventional 31 0.16 0.05 a 20 0.22 0.05 a 8 0.24 0.04 a 33 0.28 0.08 a 29 0.29 ≤ 0.03 b 6 0.10 ≤ 0.03 b 59 0.12 0.04 a 76 Label Rouge 0.12 ≤ 0.03 b 77 0.31 0.08 a 57 1.49 0.55 13 0.10 ≤ 0.03 b 39 0.12 0.04 a 63 0.12 0.06 a 38 0.15 0.06 a 78 0.18 0.05 a Organic 22 0.19 0.04 a 49 0.23 0.09 a 79 0.45 0.20 64 1.17 0.60 60 3.65 1.15 a: result comprised between the LOD (0.03 µg kg-1) and the LOQ (0.10 µg kg-1) b: result below the LOD

31

ACCEPTED MANUSCRIPT Highlights Full in-house validation by accuracy profile of a very sensitive ochratoxin A method

•

Determination of ochratoxin A in 70 samples of pig tissues

•

Three samples exceeded the Italian recommended maximum level of 1 µg OTA kg-1

•

No significant difference was found between organic and conventional productions

•

A good agreement was found between ochratoxin A levels in muscle and liver

AC

CE

PT E

D

MA

NU

SC

RI

PT

•

32