2.1. Method Validation
The developed method is selective. A Phenomenex Kinetex F5 reversed phase column (100 × 4.6 mm, 2.6 μm) produced adequate analyte retention times and separation of compounds, instead of a C18 column. Various mobile phase configurations, flow rates, and profiles were evaluated. The desired sensitivity was achieved by using water and methanol as the aqueous and organic solvents, respectively, under linear gradient elution. The analytes were loaded on the analytical column in 30% mobile phase A and eluted with 95% mobile phase B (see Section 3.6
). The optimal chromatographic conditions were flow rate 300 μL min−1
and injection volume of 20 μL. The optimized conditions have allowed for the identification of the analytes by retention times: 7.3 min for BPA, 8.0 min for E2, and 8.6 min for TT (Table 1
The best potential values for DP, EP, and collision gas flow were obtained by direct infusion of the standards switching negative/positive ESI polarity to enable the highly sensitive analysis of all analytes within a single MS assay. For better yield, negative ESI was chosen for BPA and E2 ion transitions and positive ionization for TT. The transitions monitored in MRM mode for BPA were 227.1 > 212.1 and 227.1 > 133.2 m
for its quantifier and qualifier ions, respectively, and 243.1 > 215.0 and 243.1 > 132.1 m
for its internal standard. The transitions monitored in MRM mode for 4 E2 were 271.0 > 145.0 and 271.0 > 183.0 m
for its quantifier and qualifier ions, respectively; for TT were 288.9 > 97.1 and 288.9. > 109.2 m
for its quantifier and qualifier ions, respectively. The optimized ESI and MS/MS parameters, including retention times and MRM conditions, were summarized in Table 1
The method is very sensitive. In the investigated concentration range (0.1–100 ng/mL), the calibration curves, obtained by analyzing in triplicate five standard concentrations, were linear (regression coefficient equal to 0.991 or better). The Table 2
shows the values of slope, intercept, and coefficient of determination relative to a single analyte in the different tissues. Also, the limit of detection (LOD) and the limit of quantification (LOQ) were reported. The slopes of calibration curves by standards prepared in target tissues differed at most by 4% from the slopes of calibration curves in water/methanol (matrix free), difference expressed as coefficient of variation (CV%). These results indicated there was no ion suppression or enhancement and the method was minimally affected by different matrices [43
]. These results are in agreement with those findings reported in the literature [44
] which establish that matrix effects are minor or eliminated when using LC-MS/MS. For this reason, the calibration curves were prepared in water/methanol and repeated after each set of samples.
The extraction procedure exhibited high specificity. Coupling liquid-liquid extraction (LLE) to solid phase extraction (SPE), the method was markedly selective, especially for adipose tissue, improving the consistency of chromatographic separation and prolonging chromatographic column lifetime. During the sample preparation, the lipid extract from the LLE contains many no polar compounds, such as fatty acids and phospholipids. These lipids can accumulate on the analytical columns, deteriorating the separation and providing a source of ion suppression and matrix effects in LC-MS/MS analyses [45
]. Performing SPE using specific cartridges facilitates the removal of interfering compounds, owing to the molecularly imprinted stationary phase, specific for steroids but effective for multi-residue purification, of structurally related compounds such as BPA. In the chromatograms, a lower baseline noise was observed, and no interfering peaks from the matrices were detected, proving effective sample clean-up and no co-eluting endogenous substances that could influence the ionization of the analytes.
The extraction process is efficient. By using a practical and experimental approach proposed by Matuszewski et al. [46
] to assess together matrix effect and recovery (see Section 3.7.3
), the process efficiency was calculated. In addition to the bioanalytical validation guidelines, by applying this methodology at three levels of concentration simultaneously with the accuracy and precision studies, the number of experiments to perform was minimized and at the same time required information was obtained. The percentage values reported in Table 2
ranged from 68.0% to 83.3% for five testicular tissue, from 63.7 to 70.7% for vFAT mass. The results confirmed the vFAT is a more complex matrix to analyse, but the proposed method based on liquid–liquid coupled with solid phase extraction was able to overcome matrix-related limits. In addition, the efficiency resulted very good for higher concentration levels, highlighting the possibility of using a single assay even for high analyte concentrations without the risk of saturating the cartridges and loss of sensitivity. The method is precise. The intra-day precision, expressed as RSD, was proven to be equal to or lower than 15%, whereas inter-day precision was maximum 18% (Table 3
). Accuracy, determined by comparing the mean result for five analyses to the nominal concentration value, was between 85.6 and 104.3% at all concentration levels (Table 3
To archive the matrix-independence of measures and to compensate losses during sample preparation, isotope-labeled internal standard was used in this study. In particular only BPA-d16 was chosen considering that the three analytes have a very similar structure. Conventional analytical methods commonly use internal standards or surrogates to achieve accuracy. Notwithstanding, they are time consuming and increase the cost of the analysis. Furthermore, this methodology included parameters to evaluate process efficiency, matrix effect, and ion interference and an acceptable accuracy was achieved.
2.2. Application of the Validated Method
The results demonstrated that the analytical procedure developed in this study is an effective and reliable method to extract and quantize BPA and steroids in complex matrices, such as testis and vFAT mass.
To validate the experimental procedure in real samples, the levels of BPA, E2, and TT were analyzed in testis and in vFAT mass, here used as steroidogenic tissues, targeted by BPA. Indeed, the male gonadal tissue is the primary site of E2 and TT synthesis [47
]. Estrogens are also synthesized in extra-gonadal sites, such as adipose tissue [48
], which preferentially accumulates lipophilic chemicals, as BPA [49
]. Results show that levels of BPA, E2, and TT were included in a very wide range, in both the examined tissues. For each sample (CTRL1-4 and BPA1-4), values (i.e., mean of three separate repeats ± S.D.) were expressed in ng for g of tissue (ng g−1
) and reported in Table 4
In testis from CTRL animals, BPA levels were detected below the detection limit (LOD). Such a samples showed E2 levels below the LOD, except one (1.56 ng g−1), while TT levels were really high, since these varied in a range from 215.96 to 348.57 ng g−1. In testis from BPA exposed mice, BPA levels ranged from 2.12 to 13.37 ng g−1. These samples had the highest E2 levels (from 22.56 to 228.77 ng g−1) and the lowest TT levels (from 36.76 to 83.45 ng g−1), showing that high BPA levels are related to high E2 and low TT levels.
A measurable amount of BPA, E2 and TT were detected in all the vFAT samples, also in samples from unexposed mice highlighting the difficulties in eliminating external environmental BPA- contamination. Moreover, different BPA bioaccumulation in tissues from exposed mice, in response to same BPA exposure, confirmed the need to perform a direct biological monitoring. Only measuring the BPA concentrations in target tissue, the real BPA exposure was verified and better information about the interrelationships of exposure, dose, and health effects can be provided. Differences in exposure at the equivalent dose, coupled with variations in individual susceptibility, introduce a large measure of uncertainty in the health risk assessment.
Therefore, in vFat tissues, the values ranged from 4.5 to 406 ng g−1
for BPA, from 2.46 to 201.53 ng g−1
for E2, and from 6.85 to 134.45 for TT. Three of these samples had much higher BPA and E2 levels than the other ones (on average about 12 and 20 times higher for E2 and BPA, respectively) showing that high BPA levels are related to high E2 levels. The correlation analysis confirmed that BPA and E2 levels were significantly and directly related each other (Figure 1
In agreement with the literature [18
], the data reported above replicate the findings that BPA preferentially accumulates in fat tissue and also suggest that such bioaccumulation is related to steroid levels (see BPA vs. steroid amounts in each sample). Importantly, it has been demonstrated TT/E2 ratio modulate (patho)physiology of several tissues, including testis [50
], emphasizing significance of the method here developed. Interestingly, our animals were exposed to BPA during the fetal-perinatal period and the three analytes, including BPA, were extracted and analyzed in adulthood animals, highlighting the occurrence of BPA bioaccumulation [53
]. Furthermore, the exposure was carried out primarily via pregnant/nursing mothers so that higher BPA levels detected in exposed mice agrees with previous findings that placenta [54
] and breast milk [55
] are important exposure routes.