Extraction of Benzo[a]pyrene from Moist Snuff with Water or Artificial Saliva (Part 2)

Benzo[a] pyrene and indeno[1,2,3-cd]pyrene were obtained from Sigma-Aldrich (Milwaukee, WI, USA). Other chemicals and solvents were also obtained from Sigma-Aldrich. The enzymes were obtained from MP Biochemicals (Santa Ana, CA, USA). Pure water (18.2 MΩ/cm) was obtained from a Barnstead water purification unit (Thermo Fisher Scientific, Waltham, MA, USA). Autovial PVDF filters with 0.45 μm pore size were purchased from Whatman (Florham Park, NJ, USA). Eight commercially available moist snuff samples were purchased from the U.S. market in May 2019 and stored at ambient conditions during the experiments. One additional moist snuff sample (Red Man natural fine cut) was purchased in January 2018 and kept in a freezer at −20 °C. Artificial saliva was prepared using a procedure recommended in the literature (8). This artificial saliva was prepared by dissolving in 1 L water 0.95 g KCl, 1.4 g NaCl, 0.37 g CaCl₂ × 2 H₂O, 0.84 g K₂HPO₄, 0.21 g MgCl₂ × 6 H₂O, 0.09 g urea, and 0.2 g glucose. The dissolution of the mixture generated an opaque solution. This solution was filtered after 24 h, and 1.35 g of mucin was added. The solutions of enzymes were prepared separately as follows: 1 g α-amylase was dissolved in 10 mL water, 1 mg lysozyme was dissolved in 10 mL water, and 10 mg acid phosphatase were dissolved in 10 mL water. From these solutions, 6.06 mL of α-amylase, 0.375 mL of lysozyme and 8 mL of acid phosphatase were added to 1 L artificial saliva with no enzymes. Artificial saliva was kept in a refrigerator at 4 °C, and has been utilized within a week from the time of preparation.

The high performance liquid chromatography (HPLC) analysis was performed on a 1260–1290 HPLC system from Agilent (Wilmington, DE, USA) consisting of a binary pump, an autosampler, a thermostatted column compartment, and a fluorescence detector (1290 FLD). The HPLC was equipped with a Zorbax Eclipse XDB-C18 column, 4.6 × 250 mm, 5 μm, also from Agilent. The moisture of samples was measured using a HE53 Moisture Analyzer (Mettler Toledo GmbH, Greifensee, Switzerland). Two types of shakers were used in the study. A wrist action shaker (Burrell Co., Pittsburgh, PA, USA) was used for methanol extraction of the moist snuff samples. A gyratory shaker G10 Gyrotory from New Brunswick Scientific (Edison, NJ, USA), a water bath circulator Haake A10 from Thermo Fisher Scientific (Waltham, Ma, USA) was used during the extraction of samples with water or artificial saliva. An Isotemp heating block, also from Thermo Fisher Scientific, was used during the sample preparation step.

The moisture content of the moist snuff samples was initially measured. For the HPLC analysis 500 ± 0.01 mg of moist snuff (as is) was precisely weighed in a 20-mL scintillation vial. To each vial 5 mL methanol containing 200 ng/mL indeno[1,2,3-cd]pyrene was added as an internal standard. Indeno[1,2,3-cd]pyrene is present in some moist snuff samples at levels not higher than about 10 ng/g leading to a concentration in the analyzed solution of about 1 ng/mL. This level (if present) represents 0.5% from the added compound as an internal standard and it was neglected in the calculations. The moist snuff was extracted for 30 min with agitation on a wrist action shaker.

The solution from each extract was filtered through a 0.45-μm pore size PVDF filter and placed in 2-mL screw cap HPLC autosampler vials. The HPLC separation on the Zorbax Eclipse XDB-C18 column was obtained using a gradient starting with 25% water and 75% acetonitrile for 0.5 min, then to 100% acetonitrile at 12.0 min (linear), holding at 100% acetonitrile for 6 min. At 18 min the eluant was returned to initial composition over 0.5 min and held for column equilibration for another 1.5 min (total run time 20 min). The flow rate was 1 mL/min and the column temperature was 25 °C. The detection of BaP was done using fluorescence with excitation at 378 nm and emission at 405 nm and after 15 min the excitation was changed to 370 nm and emission to 460 nm (for the detection of indeno[1,2,3-cd]pyrene). The injection volume was 20 μL. Photomultiplier gain for the fluorescence detector (FLD) was set at 13. A typical chromatogram of a moist snuff extract is shown in Figure 1. For quantitation, a calibration curve was generated using a set of nine standard solutions with BaP concentrations between 20.0 ng/mL and 0.078 ng/mL in methanol (with the internal standard at 200 ng/mL). This calibration range was selected to cover the BaP concentration in the methanol extract of moist snuff samples (that ranged between about 2.0 ng/mL and 10.0 ng/mL) as well as the concentrations recovered from the water extracts (that ranged between about 3.0 ng/mL and 15.0 ng/mL). The calibration was linear with an equation used for measuring the unknown concentrations of the form: [1]

Figure 1

Peak area normalization was obtained as the ratio of the area for the BaP peak, and the area of the internal standard (I.S.). The R² value for the linear calibration was 0.9999. The method of HPLC analysis of BaP in moist snuff samples was very similar to the one previously reported (6) and no additional validation of the HPLC procedure has been made. The extraction was assumed to be 100%, previous work (6) indicating that no detectable BaP is found using a second extraction of the samples. However, the good reproducibility (relative standard deviation RSD% = 1.17%) of the measurement of the lowest calibration point (0.078 ng/mL) indicated that this value can be used as the LOQ value.

For the evaluation of the amount of BaP extracted from moist snuff samples, 5.0 ± 0.05 g of moist snuff were precisely weighed in 250-mL flasks. 100 mL water or 100 mL artificial saliva was added to the flask. The flasks were placed in a water bath at 37 °C with rotary agitation capability and the samples were extracted for 1 h. The agitation was performed at 50 rpm.

The choice of these conditions for the extraction was made to assure exhaustive extraction in water. The water solution was placed in separatory funnels. 5 mL cyclohexane, 20 μL I.S. solution in methanol containing 10 μg/mL indeno[1,2,3-cd]pyrene, and 5 mL saturated solution of NaCl that was used for salting-out effect were added to the solution. The level of indeno[1,2,3-cd]pyrene potentially present in the moist snuff, assuming a 5% extraction efficiency in water, has been neglected in comparison with 200 ng/mL compound added as internal standard. The funnels were agitated for 15 min on a wrist action shaker. Following this step, the aqueous and cyclohexane layers were allowed to separate for about 30 min. After the layers were separated, the cyclohexane layer was placed in a 20-mL scintillation vial. The cyclohexane was evaporated at 65 °C under a mild current of air until dryness (about 20 min). The evaporated samples were re-dissolved in 1 mL methanol by placing the vials for about 1 min in a sonic bath. The generated solution in methanol was filtered through 0.45-μm PVDF filters and placed in 2-mL capped vials. These samples were submitted for HPLC analysis. This extraction procedure allowed an increase of about 50 folds in method sensitivity as compared to the procedure applied for the analysis of BaP in the moist snuff samples. The HPLC method was identical with the one previously described for the BaP analysis in moist snuff samples (6) (Figure 2).

Figure 2

Since the water/saliva extracts were processed using several steps different from simple extraction (tobacco filtration, extraction with cyclohexanol of BaP from the water/saliva solution, cyclohexanol evaporation, remaking BaP in solution of methanol), it was necessary to verify BaP recovery in the process. For this purpose, 100 mL water was spiked with two levels of BaP (and the internal standard). The water was extracted with 5 mL cyclohexane, the cyclohexane layer was separated, evaporated, and the residue re-dissolved in 1 mL methanol. The results of this recovery study are indicated in Table 1.

The recovery levels of BaP were in agreement with the level used for spiking, indicating that the method recovery is very good.

The list of moist snuff samples evaluated in this study is given in Table 2. The same table indicates the levels of moisture in the samples together with the relative standard deviations (RSD%) of the results. All samples were analyzed in duplicate. The levels of BaP measured in each sample, reported in ng/g dry material (dry weight basis), are given in Table 3. All samples were analyzed in duplicate.

The results for the 100-mL water extracts analyzed for the dissolved BaP levels are given in Table 4 as ng/g on a dry weight basis. All samples were processed in duplicate, and each duplicate was analyzed twice by HPLC. The averages from four measurements and their RSD% are given in Table 4.

The results for the 100 mL artificial saliva extracts analyzed for the dissolved BaP levels are given in Table 5 as ng/g on a dry weight basis. Similar to the case of water extraction, all samples were processed in duplicate, and each duplicate was analyzed twice by HPLC. The averages from four measurements and their RSD% are given in Table 5.

The comparison of the initial levels of BaP in 1 g moist snuff samples with the levels in 100 mL water extract (per g) are given in Table 6.

As shown in Table 6, the level of BaP extracted in 100 mL water from moist snuff ranges between 1.0% to 1.7%.

The comparison of the initial levels of BaP in the moist snuff samples with the levels in 100 mL artificial saliva extract (per g) are given in Table 7.

The results from Tables 6 and 7 indicate that only a small proportion of BaP is actually transferred from the solid material into the water or artificial saliva. However, the percentage range for extraction in water is between 1.0% and 1.7% while the extraction in artificial saliva is between 2.0% and 3.9%. Although the BaP level extracted from the moist snuff with artificial saliva remains very low, the surfactant character of artificial saliva increases to double compared to the extraction with water of BaP.