Elisabetta
Maiolini
a,
Elida
Ferri
a,
Agata Laura
Pitasi
a,
Angel
Montoya
b,
Manuela
Di Giovanni
c,
Ermanno
Errani
c and
Stefano
Girotti
*a
aDepartment of Pharmacy and Biotechnology, Alma Mater Studiorum – University of Bologna, Via San Donato 15, 40127 Bologna, Italy. E-mail: stefano.girotti@unibo.it; Web: http://bomet.fci.unibo.it/girotti/ Fax: +39 0512095652; Tel: +39 0512095660
bInter-University Research Institute for Bioengineering and Human Centered Technology (I3BH), Universitat Politècnica de València, Camino de Vera, s/n, 46022 Valencia, Spain
cRegional Agency for Environmental Protection (ARPA), Via F. Rocchi, 19, Bologna, Italy
First published on 13th November 2013
Two immunoassays, a Lateral Flow ImmunoAssay (LFIA) based on colloidal gold nanoparticle labels and an indirect competitive chemiluminescence enzyme-linked immunosorbent assay (CL-ELISA), were developed and a high performance liquid chromatography-tandem mass spectrometry (LC-MS/MS) method was optimized to assess the possible release of bisphenol A (BPA, 4,4′-isopropylidenediphenol) from different plastic baby bottles treated with simulating solutions. Coating conjugate concentration, anti-BPA antibody dilution, incubation time of the primary and secondary antibodies, and tolerance to different organic solvents were optimized to obtain the best performance of the ELISA with chemiluminescent end-point detection. The influence of different buffers on LFIA performance was also evaluated. Both methods showed good repeatability (mean CV value around 13%) and sensitivity. Reproducibility tests for CL-ELISA gave a mean CV value of about 25%. The IC50 and Limit of Detection (LOD) values of CL-ELISA were 0.2 and 0.02 ng mL−1, respectively. The LOD of LFIA was 0.1 μg mL−1. A LC-MS/MS method was also optimized. The separation was performed in a C18 column with a triple-quadrupole mass spectrometer with electrospray ionisation interface. The method showed a good linearity in the range 2 to 500 ng mL−1, with a regression coefficient of 0.998. In the simulating solutions the detection and quantification limits, calculated by the signal to noise level of 3 (S/N = 3), were 5.8 ng mL−1 and 17.4 ng mL−1, respectively. This limit of quantification was about 3 and 35 times lower than the permitted limits set by the official method CEN/TS 13130-13 (0.05 μg mL−1) and by the Directive 2004/19/EC (0.6 μg mL−1), respectively. The methods were applied to determine BPA release from baby bottles, performing repeated procedures according to EU and national regulations. The results demonstrated that no BPA migration from the tested plastic materials occurred with only one exception. The migrated amount, above the regulatory limits, was detected by all the mentioned assays.
The primary way of exposure to bisphenol A for most people is assumed to be the diet: BPA in food and beverages accounts for the majority of daily human exposure.1 The highest estimated daily intakes of BPA in the general population occur in infants and children and this creates particular concern since BPA has been identified as an important endocrine disrupting compound (EDC). It can interfere with hormonal activities by modification of biosynthesis, metabolism and elimination of natural blood-borne hormones.2 Moreover, BPA can originate environmental problems: recent research demonstrates that BPA causes inter-species mating.3
The critical period of BPA exposure is during perinatal development, when the effects tend to be permanent due to the low binding affinity of BPA to serum proteins4 and the capability of xenoestrogenic compounds to bypass the mechanisms that limit the exposure of the baby's organs and brain to circulating estrogens.5 It has been reported that exposure to low doses of BPA results in alterations in the ovary, uterus or mammary glands6 and produces a decline in reproductive capacity.7 These data confirm that during the perinatal period there is an increased sensitivity to BPA with probable consequences for babies' health.
In addition, young children have immature organ systems, high metabolic rates, and relatively low bodyweight, and go through rapid physical development, so repeated exposure to even low levels of BPA may lead to adverse health effects.7 Therefore, it is urgently necessary to have effective analytical tools available to carefully evaluate the real levels of exposure to this compound.
The U.S. Food and Drug Administration (FDA) has established a tolerable daily intake (TDI) of 0.05 mg BPA per kg body weight per day, which is also the reference dose settled by the U.S. Environmental Protection Agency (EPA). On the other hand, a specific migration limit (SML) for BPA of 0.6 mg kg−1 was set by the EU Commission in 2004.8
In September 2010, Canada became the first country in which BPA was banned.9 The European Union states outlawed the manufacture of polycarbonate feeding bottles containing the compound from March 2011, and banned their import and sale from June 2011.10 This recent BPA banning in baby bottles has forced manufacturers to employ BPA-free plastics.11 A recent study by the Joint Research Centre, European Commission,12 shows that, unfortunately, possible problems derived from the use of these alternative plastics have not been thoroughly evaluated.13
Many methods have been reported with the aim of determining BPA in water and in food: liquid chromatography tandem mass spectrometry, yeast bioassays, immunochemical assays, and HPLC with fluorescence or diode-array detectors.14–23 Immunoassays appear to comply with the sensitivity, rapidity, and low cost requirements for effective monitoring activities.24,25
Here we report on the development of a colloidal gold-based lateral flow (LFIA) assay, on the optimization of an ELISA with chemiluminescent detection (CL-ELISA), both based on an indirect competitive immunoreaction, and on the optimization of a High Performance Liquid Chromatography Tandem Mass Spectrometry (LC-MS/MS) method for BPA analysis. The three methods were applied to analyze the possible BPA migration from baby bottles treated to simulate the real, repeated conditions of use, taking into account the food characteristics and the sterilization and heating procedures usually involved. This study was performed according to the statement: “heating bottles or pouring hot liquids into bottles, the presence of acidic or basic foods and beverages have all been shown to increase the rate of BPA leaching from bottles”.26
The main buffering solutions employed were: 0.05 M carbonate–bicarbonate buffer , pH 9.6; 10 mM PBS , pH 7.4 (10 mM Na2HPO4, 2 mM KH2PO4, 137 mM NaCl, 2.7 mM KCl); PBST (PBS containing 0.05% Tween20); PBSG 1 (10 mM PBS, 0.5% Fish Gelatine); PBSG 2 (20 mM PBS, 1% Fish Gelatine), and 0.2 M BB (borate buffer: 50 mM Na2B4O7, 200 mM H3BO3, pH 8.5).
In order to simulate the migration of bisphenol A from baby bottles under different conditions several assays were carried out using liquids simulating different kinds of food, as indicated in the European Commission Directives 93/8/EEC and 97/48/EEC.27,28 Three different solutions were used:
- Simulating solution A: distilled water. It reproduced the extraction capability of aqueous food with pH > 4.5.
- Simulating solution B: water–acetic acid 97:3 (v/v). It reproduced the extraction capability of acidic food with pH < 4.5.
- Simulating solution C: water–ethanol 50:50 (v/v). It reproduced the extraction capability of fatty food (particularly milk).
The migration tests were conducted in order to reproduce the conditions of real home-use during the disinfection and meal preparation steps, taking into account the indications reported in the leaflets of the different bottles which recommend a 5 minutes boiling in water prior to the first use. To this end, the bottles of the same material (PP or PC) were divided into two groups of 6 bottles each: one group was subjected to the initial sterilization by 5 minutes boiling whereas the other one was not. Each group of 6 bottles was further divided into two groups of 3 bottles made by the same material and each bottle was filled with one of the simulating solutions A, B, or C. One group was treated by microwaves and the second one by water bath heating for 6 consecutive cycles, replacing the simulating solutions after every cycle (Fig. 1). Only the simulating liquids obtained after the third and sixth migration cycles were analyzed, as indicated by the Italian legislation.29 Further tests were carried out under conditions of stress simulated by the two cycles of sterilization, the first one by heat and the second one by cold sterilization. After each sterilization cycle 3 migration cycles in a water bath or microwave oven were applied. Time and temperature exposure conditions complied with the EU legislation27 and the instructions for use of each bottle.
Fig. 1 Scheme of the procedure to induce migration of BPA from baby bottle simulating conditions of repeated use. |
For polycarbonate bottles (containing BPA): 30 min exposure at 100 °C. For bottles made of PP: 5 min exposure at 50 °C (Fig. 1, step A, right).
The optimized calibration curves were performed in the 0.001–1000 ng mL−1 range. The OVA-BPAB conjugate (0.2 μg mL−1) in 0.05 M carbonate–bicarbonate buffer, pH 9.6 (100 μL per well) was added into the black polystyrene high-binding microplates (Costar, Cambridge, USA) and incubated overnight at 4 °C for coating. Plates were washed 3 times with 200 μL per well of PBST using a microplate washer (WellWash 4, Labsystem, Finland) and the uncoated sites were blocked by adding 200 μL per well of PBSG 1×. The plates were incubated at room temperature (RT) for 60 min with shaking and washed again 3 times as described above. For the competitive reaction 50 μL per well of 0.2 μg mL−1 of BPAB-11 antibody in PBSG 2× and 50 μL per well of standard solution or sample extract were added and incubated for 60 min at RT without shaking. In order to avoid the absorption of BPA on plastic surfaces only glass containers were used to manipulate the standard solutions (glass vials, Hamilton syringes, etc.). After another microplate washing, 100 μL per well of rabbit anti-mouse IgG-HRP diluted 1:2000 in PBSG 1× were added and incubated for 60 min at RT with shaking. Finally, after a new washing step, 100 μL of luminescent mixture (45 μL of 1 mM luminol, 10 μL of 0.5 mM p-iodophenol, 9.8 mL of 0.2 M borate buffer, pH 8.5 and 100 μL of 1 mM hydrogen peroxide) were added to each well and the luminescence emission at 425 nm was immediately recorded using a Victor 1420 luminometer (Wallac-Perkin Elmer, Waltham, MA, USA) and expressed as Relative Luminescence Units (RLU).
The chemiluminescent emission values (Ex) were normalized between 100% (the emission of a blank control, E0) and 0% (the emission of a sample with an excess of standard compound, Eexcess) according to the expression:
%B/B0 = 100(Ex − Eexcess)/(E0 − Eexcess) |
The results, expressed as “%B/B0”, were the ratio between B, the emission of “X” ng mL−1 standard solution and B0, the emission of a 0 ng mL−1 standard solution.30 The normalized values were then mathematically fitted, using the Sigmaplot® software (SPSS) version 8.0, to the four parameter logistic equation:
y = {(A − D)/[1 + (x/C)B]} + D |
The detection of BPA was carried out by reverse phase chromatography using a Supelco Discovery® C18 (150 × 2.1 mm, 5 μm) column from Supelco (Milan, Italy) thermostated at 30 °C, with phase eluent consisting of demineralized water MilliQ (mobile phase A) and a solution of 0.1% (v/v) of ammonia in methanol (mobile phase B) at a constant flow of 200 μL min−1. The elution conditions started from a mixture of 90% of phase A, followed by a linear gradient of 5 min up to 85% of phase B, a plateau of 10 min, return to initial conditions in 2 minutes and finally 5 min of column reconditioning. The injection volume was 50 μL. The total elution time was 22 min, with a retention time of 11 min for bisphenol A.
Compound | Ion mode | Cone voltage (V) | Transition quantification | Coll. E (eV) | Transition confirmation | Coll. E (eV) | Dwell (s) |
---|---|---|---|---|---|---|---|
BPA | ESI | 31.0 | 227 > 212 | 20.0 | 227 > 133 | 50.0 | 0.3 |
d16-BPA | ESI | 31.0 | 241 > 142 | 20.0 | — | — | 0.3 |
Standard BPA conc. (ng mL−1) | Repeatability CV (%) (n = 6) | Reproducibility CV (%) (n = 6) | Recovery (%) (n = 3) |
---|---|---|---|
0.02 | 22.3 | 35.5 | 84.3 |
0.2 | 15.7 | 28.9 | 86.1 |
2 | 12.5 | 25.6 | 85.4 |
7 | 10.6 | 19.8 | 95.6 |
50 | 9.7 | 22.1 | 91.7 |
100 | 8.2 | 18.2 | 104.2 |
The specificity of monoclonal antibody BPAB-11 had previously been evaluated by performing competitive assays with several compounds structurally related to bisphenol A.17 Significant cross reactivity values were observed only with the closely related compounds 4,4-ethylidenebisphenol (6–19%) and 4-cumylphenol (20–74%).
The quantification transitions chosen were 227 > 212 for the bisphenol A (due to the loss of a methyl group34) and 241 > 142 for the deuterated bisphenol A (due to the same molecular rupture), the confirmation transition detected of the bisphenol A was 227 > 133 (due to the loss of a phenol group). Table 1 summarizes the setting parameters of the spectrometer and Fig. 4 shows the spectra acquired during scanning in a range from 0 to 400 amu (atomic mass unit) for the identification of the molecular ion m/z 227 and in a range from 100 to 400 amu for the fragments of transition m/z 212 and 133 for the bisphenol A and m/z 142 for the d16 BPA.
The developed method showed a very good linearity in the 6–500 ng mL−1 range, with a correlation coefficient of 0.998. The detection limit in the simulating solutions calculated by the signal to noise level of 3 (S/N = 3) was 5.8 ng mL−1, 9 times lower than that stated in the official method UNI CEN/TS 13130-13 [Ref. 35] (0.05 μg mL−1) and about 100 times lower than the limit set by the 2011/10/EU Directive for the BPA migration in food (0.6 μg mL−1).8 The quantification limit, calculated according to the same criteria (S/N = 10), resulted in 17.4 ng mL−1, a value about 3 and 35 times lower, respectively, than the above mentioned permitted limits.
We determined some analytical parameters such as the repeatability, assessed at 450 ng mL−1 (n = 10) with a CV of 6%, and the reproducibility of a 121 ng mL−1 standard solution (n = 6), obtaining in this case a CV of 12.5%.
Method | BPA content (μg mL−1) | Number of determinations (n) | SD (%) |
---|---|---|---|
CL-ELISA | 4.5 | 3 | 12.5 |
LC-MS/MS | 3.3 | 3 | 9.8 |
LFIA | + (up to 1000) | 3 | — |
All other samples were below the limit of quantification of the quantitative methods and then extensively below the admitted limits set by regulations, showing that the migration of BPA from baby bottles still containing this compound is a quite rare eventuality, even after repeated application of sterilization and heating treatments. Moreover, one sample is not enough to confirm that the 5 min boiling treatment can influence the migration process, at least in the case of subsequent treatment by microwave. These results are in accordance with recently reported studies12,13 that examined the BPA release from 277 baby bottles made of several types of plastic (polycarbonate, polyamide, polyethersulphone, polypropylene, silicone, and Tritan™) and purchased from 26 European Union countries, Canada, Switzerland and the USA.
It is surely important to underline that by comparing the data obtained from the LC-MS/MS analysis with those of the two immunoassays not one false positive result from the CL-ELISA or the LFIA was observed, supporting the reliability of these methods.
The new, simple and fast LFIA has been revealed to be a very interesting method for the semi-quantitative determination of bisphenol A. The detection limit was very good for such a type of portable assay and the LFIA format was the only one suitable to solve the reproducibility problems of ELISA, connected with the property of BPA to absorb on plastic surfaces: the run can be entirely performed in a glass tube.
Bisphenol A is currently detected by chromatographic separation (LC) often coupled with mass spectrometry.14,34,36 To ascertain the reliability of the developed immunological methods we set up this simple LC-MS/MS method which was able to determine BPA at concentrations below the lower limits set by the present legal rules and with a LOQ lower than the LC-DAD official method. This LC-MS/MS method does not require any derivatization procedure or pre-concentration step. The significant reduction in the time per analysis has been further increased by employing the quicker separation on gradient instead of the isocratic one. The very high sensitivity achieved makes this assay a suitable control method to determine the absolute content of BPA and to reveal the migration of very low amounts of BPA from plastics.
This journal is © The Royal Society of Chemistry 2014 |