Fumonisins are mycotoxins produced by Fusarium, mainly F. verticillioides
]. These compounds are found worldwide, sometimes at relatively high levels in human food and animal feed [1
]. Because of their fungal origin, not only one metabolite is produced, and 28 fumonisin analogs have been identified until now [1
]. The most widely studied fumonisins belong to the “B” family (FB), FB1 being the most abundant, the other being FB2, FB3, and FB4 [1
]. Other fumonisins produced by Fusarium found in food and feed are fumonisin A (FA), fumonisin C (FC), fumonisin P (FP) and hydrolyzed and partially hydrolyzed fumonisins (HFB). Although FA, FC, FP, and HFB have been shown cytotoxic and inhibit sphingolipid synthesis, FB are the most abundant and the most toxic compounds [1
]. Accordingly, the provisional maximum tolerable daily intake (PMTDI) and recommended guidelines on fumonisins in food and feed have been established based on the cumulated intake of FB [1
After their administration in animals, FB are poorly absorbed and rapidly excreted [8
]. Only a very small amount of the administered dose is found in plasma, and the metabolism of FB appeared to be weak [8
]. Although HFB, FA, and N-carboxymethyl FB have been found in the liver and feces of different species, the mechanism of their formation is not well understood and their contribution to the overall toxicity of FB is considered to be insignificant compared to that of the parent compound [8
Taking into account the levels and occurrence of FB in raw materials, their poor absorption in animals and their weak level in milk, human exposure to FB through consumption of animal products and products of animal origin is considered to be negligible [1
]. However, to date, no data are available on levels of FB in the muscles of poultry that can tolerate high levels of FB in their feed [2
]. Also, as multiple contamination by toxins produced by Fusarium
is common in poultry diets, and some fusariotoxins are known to change xenobiotic and nutriment absorption, concomitant exposure to several toxins could change the level of FB in tissues [10
]. Specifically, deoxynivalenol (DON) is known to affect the intestinal barrier function in several animal species, which could modify the bioavailability of xenobiotics [16
]. However, chronic exposure to DON appeared to have no influence on the oral bioavailability of a single dose of FB1 [23
]. Concerning zearalenone (ZEN), a study in broiler chickens showed that dietary ZEN improved nutrient digestibility, suggesting FB bioavailability could change during concomitant exposure to FB and ZEN [24
Several methods of analysis of fumonisins have been developed in plant and biological samples, however the UHPLC-MS/MS methods have been shown to have the highest sensitivity and specificity [25
]. The objectives of these methods varies with the samples analyzed. Indeed, whereas the main objective of UHPLC-MS/MS analysis of food and feed is usually to detect multiple mycotoxins, the main objective of analysis using biological matrices is sensitivity, especially for FB, whose level in sample is generally. Most UHPLC-MS/MS methods used for the analysis of biological samples involve precipitation of proteins with organics solvents, liquid–liquid extraction and solid-phase extraction before LC-MS analysis [26
]. The columns usually used to purify the samples are of the SAX or C18 type, whereas immunoaffinity (IA) columns are rarely used in UHPLC-MS/MS analysis, except to clean some plant samples [31
]. IA columns have been used with HPLC analysis and fluorescent detection for quantitation of FB1 in liver, but the relatively low sensitivity of fluorescence detection did not enable quantitation of other FB than FB1, nor quantitation of FB1 in muscles [33
So, although a PTDI for FB1, FB2 and FB3 of 2 µg/kg body weight has been established, no data are available on the level of FB in muscle of avian species. Because broilers and turkeys are major sources of meat and because these species can tolerate high levels of FB in feed, assessment is of special interest. Also, because of the multiple occurrence of fusariotoxins in feed, understanding the consequences of multiple exposure on the level of FB in tissues is also of interest. To this end, we first developed an UPLC-MS/MS method that enables simultaneous quantitation of FB1, FB2, and FB3 in muscle and liver. We then applied the method to analyze samples that had been taken as a part of two toxicological studies in chickens and turkey fed with FB alone, and fed with FB in combination with DON and ZEN [34
3. Material and Methods
3.1. Tissue Samples
Tissue samples were obtained from chickens and turkeys fed with experimental diets to investigate the effects of fusariotoxins on health [34
]. The experimental protocols were approved by the French Ministry of Higher Education and Research and registered under number 02032.01. Briefly, ground cultured toxigenic Fusarium
strains were incorporated in corn–soybean diets formulated to best meet the nutritional needs of the animals. The control diets (Control) were free of mycotoxins, the fumonisin (FB) diets were formulated to contain 20 mg FB1+FB2/kg, and the fusariotoxin diets (FBDONZEN) were formulated to contain 20, 5, and 0.5 mg/kg of FB1+FB2, DON and ZEN, respectively. Each of the experimental diet was distributed ad libitum to 14 broilers from the 1st to the 35th day of age and to 14 turkeys from the 55th to the 70th day of age. After a starvation period of 8 h, the animals were killed by exsanguination after stunning by electrocution. The liver and the breast muscles were collected and stored at −80 °C until analysis. No signs of toxicity were observed, and only slight differences were found between groups on performance, organ weight, histopathology, intestinal morphometry and the number of goblet cells, oxidative damage, sphingolipid metabolism, and male reproductive toxicity [34
3.2. Fumonisins, Reagents and LC-MS/MS Conditions
All reactive and reagents were purchased from Sharlab S.L. (Sentmenat, Spain). Standard solutions of FB1, FB2, FB3, U-[13
]-FB1, and U-[13
]-FB2 with certified concentrations of each analyte were purchased from Biopure (Tulln, Austria). The UPLC MS/MS system, including the software used to treat the chromatograms, was purchased from Agilent (Santa Clara, CA, USA). The UPLC system was a 1260 model composed of an automatic injector, a degasser, and a binary pump. A Poroshell 120 column (3.0 × 50 mm, 2.7 µm) was used for the separation step. A 6410 triple quad was used for detection after positive electrospray ionization. Source parameters were adjusted as follows: the temperature of the gas was set at 300 °C, gas flow at 10 L/minute, nebulizer was 25 psi, capillary voltage was 4000 V. Table 1
lists the optimized MRM conditions used for LC-ESI-MS/MS analysis for each analyte. The most abundant transition was chosen for MRM quantitation, while two other transitions were used as qualifiers for FB1, FB2, and FB3. Only one transition was used for qualification of C13FB1 and C13FB2.
The mobile phase was composed of a mixture of methanol (solvent A) and water (solvent B), each containing 0.1% formic acid (v/v), and was delivered at a flowrate of 0.3 mL/min. Solvents A and B were in the same proportion at the beginning of the run then a gradient of elution was introduced to reach 95% of A and 5% of B at 5 min. Return to the initial conditions was achieved at 8 min, then 4 min of washing was done before a new run was performed. The volume of injection was 10 microliters.
3.3. Analysis of Standards Solutions and Efficiency of Immunoaffinity Columns
Certified standards solutions were diluted in acetonitrile/water (1:1) to obtain working solutions containing mixture of FB1, FB2, and FB3 at 500 (FB500), 100 (FB100), and 20 ng/mL (FB20) and mixtures of C13FB1 and C13FB2 at 500 (C13FB500) and 100 ng/mL (C13FB100). Variables volumes of working solutions were evaporated to dryness. Dry residue was solubilized in 200 µL of mobile phase composed by a 50/50 mixture (v/v) of solvent A and B. Concentrations of each analyte in the injected solutions were 0, 2, 10, 50, and 100 ng/mL A quadratic fit of measured signal (y-axis) vs concentration (x-axis) was used. Accuracy was calculated at each concentration and was considered as acceptable for a relative standard deviation (RSD) of 20%.
The recovery rates of standards solutions passed through the IA columns were measured at different concentrations. Variable volumes of working solutions were solubilized in a final volume of 10 mL of ACN/MeOH/PBS (5:5:90, v/v/v) and passed through a FUMONIPREP column (R. Biopharm Rhone Ltd., Glasgow, Scotland) according to the manufacturer instructions. Columns were washed by 10 mL of 2mM pH 7.3 saline phosphate buffer (PBS), and eluted with 1.5 mL of methanol followed by 1.5 mL of water. The eluate was collected, evaporated to dryness, and stored at −20 °C until analysis. Before analysis, the dry residue was suspended in 200 µL of mobile phase composed by a mixture of solvent A and B (50:50, v/v) and sonicated for 5 min. Solubilized residue was centrifuged 10 min at 3000× g, the supernatant was collected and placed in chromatographic vials. Expected concentrations of FB1, FB2, and FB3 were 0, 2, 10, 50, and 100 ng/mL while expected concentrations of C13FB1 and C13FB2 were 62.5 ng/mL
3.4. Treatment of Tissue Samples and Determination of the Recovery Rates
Five g of muscle were homogenized in 5 mL of distilled water with an Ultra Turrax. Then 25 mg of NaCl, 25 µL of a working solution C13FB500, and 5 mL of acetonitrile/methanol (1:1) were added. Livers were prepared in the same conditions except 1 g of tissue was homogenized in 2mL of water and 2 mL of acetonitrile/methanol was added. Homogenized samples were placed on a stir table at 300 rpm for 2 h and centrifuged for 15 min at 3000× g. The supernatant was collected, 8 mL of hexane was added, and the mixture was vortexed for 30 seconds. The organic phase (upper) and the aqueous phase were separated by 15 min of centrifugation at 3000× g. For muscles samples, 5 mL of the aqueous phase were collected and 20 mL of 2mM pH 7.3 PBS were added. For the liver samples, 2 mL of the aqueous phase were collected and 8 mL of PBS were added. Extracts solubilized in PBS were passed through a FUMONIPREP column as previously described.
Matrix interactions were measured on tissue samples obtained from animals not exposed to FB in their diet over a period of at least 15 days. Muscle and liver were extracted and purified as previously described except the lack of fortification with C13FB500. Variable volumes of working solutions were added to the dry residue and evaporated to dryness. The dry residue was solubilized in 200 µL of mobile phase as previously described. Expected concentrations of FB1, FB2, and FB3 were 0, 2, 10, 50, and 100 ng/mL while expected concentrations of C13FB1 and C13FB2 were 62.5 ng/mL.
The recovery rates of standards solutions of FB were measured in fortified blank muscle and blank liver samples obtained from birds fed the mycotoxin-free diets. Tissue samples were prepared as previously described, fortified with 25 µL of IS500 and variable volumes of FB500, FB100, and FB20 to obtain supplementation levels equivalent to 0, 0.25, 1, and 5 ng/g of FB1, FB2, and FB3. Because FB1, FB2, and FB3 concentrations in diets are different, two other assays were performed to obtain final FB1, FB2, and FB3 concentrations of 25, 5, and 5 ng/g, and 100, 5, and 5 ng/g, respectively. In all the assays, the concentrations of C13FB1 and C13FB2 were 2.5ng/g in muscles and 12.5 ng/g in liver. The intra-day repeatability (n = 5) and inter-day reproducibility (5 days) of the whole method were calculated for C13FB1 and C13FB2 on muscle and liver spiked samples and expressed by the RSD of the concentrations measured. The LOQ was defined as the lowest concentration of a sample that can still be quantified with acceptable precision and accuracy (bias). The acceptance criteria for these two parameters were 20% RSD for precision and ± 20% for bias [41
3.5. Statistical Analysis
The calibration curves obtained after passage of standard solutions on IA column and the calibration curves done to assess matrix effect in each species were compared to standard calibration curves using two-tailed paired t-test. Recovery rates, matrix effects and species effects were compared using ANOVA. When a significant difference was found (p < 0.05) a complementary comparison of mean was done using the Kruskall-Wallys test. Groups that are statistically different (p < 0.05) are identified by different letters.