Chemicals
The PFOA (≥ 98%) for dust coating experiments was purchased as solid standard from Sigma Aldrich (St. Louis, MO, USA). Analytical native standard of PFOA (perfluoro-n-octanoic acid), mass-labelled extraction standard served as internal standard (IS), perfluoro-n-[1.2.3.4- 13C4]-octanoic acid), and mass-labelled injection standard served as recovery standard (RS), perfluoro-n-[ 13C8]-octanoic acid), were purchased from Wellington Laboratories (Guelph, Ontario. Canada). Ammonium acetate, methyl-tert-butyl ether (MTBE) (≥99.8%), tetrabutylammonium bisulfate (TBA) with a purity of ≥99% and methylpiperidine (1-MP, 99%) were purchased from Sigma-Aldrich® (St. Louis, MO, USA). OASIS® WAX (6 mL, 150 mg) was from the Waters Corporation (Milford, MA, USA). Formic acid (>95%), methanol HPLC-grade (≥99.8%), methanol LC-MS-grade (≥99.9%), acetonitrile HPLC-grade (≥84 99.9%) and ammonium solution (25%) were purchased from Fisher Scientific (Pittsburgh, PA. USA).
Respirable fraction of house dust
The house dust was obtained from vacuum cleaner bags collected from 32 residential homes in the region of Stockholm, Sweden. The respirable fraction of house dust was retrieved by a sieving process of which all 32 vacuum cleaner bags were pooled. The processing of dust, to a respirable particulate size fraction have been described in detail by Gustafsson et al. 2018 (Gustafsson et al. 2018). Briefly, the dust was sieved through six stages of plane woven steel meshes, followed by passage of a cyclone, after which the dust was collected in a filter bag. The dust was once again sieved through a twilled woven steel, yielding the respirable fraction. The sieving process was performed under constant airflow and mechanical deagglomeration.
Sorption of PFOA on house dust in the gas phase
To achieve a sufficient concentration of PFOA on the house dust, to ensure detectability in the tissue samples, a near saturated vapor concentration procedure was adopted to adsorb/absorb PFOA on the house dust. Ten grams of solid pure perfluorooctanoic acid was placed in a glass jar bottle, capped with a lid, shaken and stored at room temperature (25±0.1°C) to allow vaporization and pre-equilibration of PFOA on the walls of the glass jars. Following equilibration of PFOA in the air and on walls of jars, a portion of 5±0.03 g dust was placed in a polypropylene (PP) round plate inside each of four jars. These were capped and kept under the controlled room temperature and atmospheric pressure over time, mimicking a residential indoor environment, but with a near saturated vapor of PFOA. Repeated samples of dust were withdrawn from the jars over time and analyzed for the PFOA content. This procedure was repeated over time for up to 6 weeks until a stable equilibrium concentration of PFOA was attained adsorbed to the dust. The long equilibration time in the near saturated gas-phase of PFOA allowed for an even adsorption/absorption on the dust particles in the testbed.
During the adsorption process, the concentration of PFOA adsorbed to the dust was periodically measured after a rotatory mixing of dust samples with a tube rotator or 24 h prior to the extraction procedures and instrumental analysis. Preliminary trials showed that this rotatory mixing prevented the formation of pockets of condensed PFOA at the outer edges of the sieved dust facing higher air concentrations of PFOA. The procedure resulted in a uniform adsorption over the entire dust sample on the PP plate. Measurements were repeatedly performed until a saturation of PFOA was reached on the dust. At the time when the concentration was within 10% of the previous measurement the dust was considered as saturated. The dust was stored in a -20°C freezer until the animal exposure experiment. Four replicated dust adsorption tests were prepared identically in this manner to verify the reproducibility of the dust adsorption experiment and homogeneity of the PFOA adsorbed dust samples.
Animals
This study was conducted in accordance with a protocol approved by the Animal Committee of ethics in Linköping, Sweden according to Directive 2010/64/EU.
Male Rat RccHan:WIST, approximately 22 weeks of age (450 - 575g), were purchased from Envigo, Venray, Netherlands. The rats were housed five each in clear polycarbonate cages (20×25×47cm) containing enrichment such as nesting material, chew sticks and tunnels. Water from the municipal tap and extruded rodent diet 2016 Teklad global, 16% protein (Envigo. Venray. Netherlands) were provided ad libitum. Animals were housed in a facility that maintained an average temperature of 23.2°C (Min: 21.6 – Max:23.9), average humidity of 51.1% (Min:25.4 – Max:60), and a 12 h light :12 h dark cycle.
Experimental design
A total of 10 animals were used in the study. Four rats in each group were exposed by either inhalation- or gavage, respectively. Two rats were unexposed and used as experimental control animals. For both exposed groups, the blood samples were collected prior to exposure (0h) and then at 3, 6, 24 and 48 h post exposure. The blood was kept on ice after collection. Blood samples were collected from the control rats at the time of termination. The blood samples were centrifuged at 2000 g for 5 min and the plasma samples were stored at -20°C until the analyses. At the time of termination, 48 h after exposure, the rats were exsanguinated under isoflurane and oxygen anesthesia. Bronchoalveolar lavage fluid (BALF) was collected with 5 mL of sodium chloride and stored in a freezer at -20°C. The lungs, liver and kidney were removed and frozen in liquid nitrogen followed by freezer storage (-20°C) until analyses. All plastics that were used were made of PP or high-density polyethylene to avoid loss of PFOA due to adsorption to the plastics.
Aerosol generation and aerodynamic particle size distribution
An aerosol was generated with the PreciseInhale® platform (Inhalation Sciences Sweden AB, Stockholm, Sweden) and particle size distribution of the house dust was measured with a nine-stage Marple cascade impactor (Marple and McCormack, 1983). The dust was aerosolized batch wise into a 300 mL holding chamber of the PresiceInhale® and then pushed out at a flow rate of 330 mL/min. Prior to aspiration into the impactor, the aerosol was diluted into a continuous airflow of 2 L/min. The mass of dust deposited on the nine stages in the impactor were used to calculate the Mass Median Aerodynamic Diameter (MMAD) and the Geometric Standard Deviation (GSD). The method was adapted from Selg and co-workers (Selg et al. 2010; Selg et al. 2013) and the aerodynamic characterization of the respirable fraction of the pooled house dust used in this study has previously been described by Gustafsson et al. 2018 (Gustafsson et al. 2018).
PFOA dust inhalation exposure
The house dust was aerosolized with the DustGun powder generator and then delivered to the rats with the PreciseInhale dispensing system. In order to determine the system settings for reaching the target dose of dust, pre-exposure filter experiments were performed with the PreciseInhale system. The inhaled mass of the dust collected using an in vitro filter test system for rats exposed in vivo by intratracheal intubation, was calibrated against the optical signal from a Casella Microdust Pro light dispersion instrument (Casella CEL, Inc., Buffalo, NY). An inhaled dose of approximately 0.5 mg of dust per animal was decided; to compare to 0.5 mg for the gavage. The achievable dose of 0.5 mg house dust inhaled resulted in a deposited dose of 0.26 mg per rat, based on a calculated deposition fraction of 0.51 as determined with the MPPD model for the particle size distribution of the house dust (Asgharian et al. 2002). Because of a time limitation on how long the animals could be kept intubated, the previously achieved deposited dose by gavage could not be fully reached via inhalation. However, for organic inhalants such as PFOA, dose normalization between adjacent exposure levels can be accurately achieved with retained accuracy of the pharmacokinetics (Malmlöf et al., 2019).
The rats were anesthetized by administrating a premedication of 0.05 mg/mL atropine (50 µg/kg bw) subcutaneous followed by a cocktail of fentanyl (10-15 µg/ kg bw) and Dormitor Vet. (0.21- 0.30 mg/kg bw) (50:50) that was administrated gradually in the tail vein intravenous until anesthesia was achieved. The rats were intubated with a stainless-steel catheter (outside diameter 2.02 mm, inside diameter 1.67 mm, length 6 cm) using a laryngoscope. The intubated rat was placed on heating pad in supine position on an adjustable table and connected to the PreciseInhale system. Each animal was monitored for 5 minutes before aerosol exposure to ensure stable spontaneous breathing. Four animals were exposed to 9-10 exposure shots of dust during an exposure time of approximately 10 minutes. The generated aerosol was drawn from the aerosol holding chamber past the breathing animal at a superimposed flow rate of 340 mL/min, which was close to the optimal relation between ventilation rate and the superimposed flow rate for wasting a minimum of test substance, yet preventing rebreathing of exhausted aerosol (Moss et al. 2006). After the inhalation exposure the rats was injected intramuscularly with Naloxon (0.08-0.1 mg/kg bw) for a fast recovery from anesthesia.
PFOA dust oral gavage exposure
A suspension of 0.5 mg of dust was prepared in 2 mL tap water. The rats were gavaged by a single dose of dust with PFOA. After the delivery of dust, the tube was rinsed with an additional volume of 1.5 mL of tap water which was also delivered to the rats.
Chemical analysis
Gas phase PFOA sampling and analysis
The gas phase of PFOA was collected with OSHA Versatile Sampling (OVS) tubes with XAD-2 resin (SKC). Samples were collected at a flowrate of 0.04 L/min using vacuum sampling pumps. Collection times were set for 5 min typically for a 200 mL air sample. PFOA from the OVS sampler was extracted prior to analysis into two fractions with each fraction placed into individual PP tubes. Fraction A consisted of the glass or quartz fiber filter and first section of XAD resin beads and the first polyurethane foam (PUF) filter. Fraction B consisted of the second section of XAD resin beads and the back PUF filter. All samples were spiked with 50 µL of 1 µg/mL C13 solution, the surrogate standard. Each section was placed in PP tube to which 5 mL of HPLC-grade methanol was added. The tube was shaken for 60 min on a mechanical flatbed shaker set at ∼2 cycles per sec. Approximately 4.5 mL of each extract was transferred to a pre-baked amber glass vial by filtering through a Gelman GHP Acrodisc (Pall Gelman Laboratory, Ann Arbor, MI, USA.). Each extract of consisting of 200 µL was transferred to an autosampler vial with PP insert. After that, 10 µL of recovery standard was added to the autosampler vial for PFAS analysis using LC-MS/MS.
Respirable fraction of dust
Internal standard (50 µL) was added to the sieved dust (50 mg) and was left to equilibrate overnight. The sample was extracted with methanol (2 mL) with 30 sec vortex and 15 min sonication in-between. The extract was centrifuged (3000 rpm) for 10 min and the supernatant transferred into a new tube. The procedure was repeated once with methanol (1 mL). The combined extract for PFOA analysis was cleaned up by diluting it in 25 mL MilliQ water and the sample was adjusted to pH 10 with ammonium hydroxide. Hexane (4 mL) was added to the water, separated by centrifugation (3000 rpm) and the hexane phase was discarded. The pH of the water was adjusted to pH 4 by adding formic acid. The extract was added to a WAX cartridge (Oasis WAX 150 mg 6cc. Waters), washed with ammonium acetate (pH 4) and tetrahydrofuran (Biosolve)/methanol (3/1), and the analyte eluted with methanol 0.1% ammonium hydroxide. The eluate was evaporated to dryness and reconstituted in 40% methanol in ammonium acetate (2 mM).
Plasma and lung lavage
The procedure for extraction of plasma samples followed published method (Yeung et al. 2009) with some modifications. Briefly, 10 µL of mass-labelled internal standard was spiked in to the 50 µL diluted plasma samples (in MilliQ water); they were mixed with 1 mL of 0.5M TBA solution in a 15 mL PP tube. After mixing, 5 mL of MTBE was added, and the mixture was shaken for 20 min at 250 rpm. The organic and aqueous layers were separated by centrifugation at 3000 rpm for 15 min. The organic layer MTBE (4 mL) was transferred to a new PP 15 mL tube. The extraction was repeated twice with 5 mL of MTBE was removed each time. The three extracts were combined in the second PP tube. A 1 mL of methanol was added to the final extract before it was concentrated to 1mL under nitrogen. The extract was further concentrated to 200 µL in a LC vial, and mass-labelled recovery standards were spiked into the vial. Aqueous ammonium acetate (2 mM, 300 µL) was added to the vial for PFAS analysis using LC-MS/MS.
Tissues
All tissue samples were homogenized using a Tissue Tearor (BiospecProducts) with tissue (1 g wet weight) and 1% potassium chloride solution (0.2 mL); the homogenate was then spiked with 10 µL of mass-labelled internal standard before the ion-pairing extraction procedure as described above for the plasma samples.
Extractable organofluorine (EOF) analysis
Plasma and tissue samples taken at 48 h after exposure were subjected to EOF analysis. Extraction procedure followed the procedure described above with the exception that no mass labelled internal standards were spiked to the samples before extraction. The EOF content of the 1 mL extract was analyzed by combustion ion chromatography (CIC). Part of the extract (200 µL) was transferred to a LC vial with the addition of 10 µL of mass-labelled standards; after that aqueous ammonium acetate (2 mM, 300 µL) was added to the vial for PFAS analysis using LC-MS/MS. The reported EOF concentrations as well as the concentration of PFOA for mass balance analysis were not recovery-corrected.
Instrumental analysis of PFOA
PFOA was analyzed using a Waters Acquity UPLC coupled to a Waters Xevo TQ-S triple quadrupole mass spectrometer operating in negative ion mode for electrospray ionization. A Waters Acquity UPLC BEH C18 column (1.7 µm, 50 × 2.1 mm, Waters) was heated to 50°C with a flow rate of 0.5 mL/min. A 10 µL extract aliquot was injected onto the column with mobile phases consisting of 2 mmol/L ammonium acetate in a mixture of MeOH and water MeOH 30/70 (v/v) (A) and 2 mmol/L ammonium acetate in MeOH (B). Details of the LC-MS method are presented elsewhere (Aro et al., 2021).
Instrumental analysis of extractable organic fluorine (EOF)
Levels of EOF in the sample extracts were determined with a combustion ion chromatography (CIC) system consisting of a combustion module (Analytik Jena, Germany), a 920 Absorber Module and a 930 Compact IC Flex ion chromatograph module (both from Metrohm, Switzerland). An ion exchange column (Metrosep A Supp 5 – 150/4.0), with carbonate buffer (64 mmol/L sodium carbonate and 20 mmol/L sodium bicarbonate) as the mobile phase, were used for the separation of anions; the absorber solution was water. The sample extract (100 µL) was set on a silica boat via an autosampler and placed into a furnace at 900–1000°C. The combustion of the sample in the furnace converted organic fluorine and inorganic fluorine into hydrogen fluoride (HF), which was then trapped by MilliQ water. The fluoride concentration in the solution was analyzed using ion chromatography. A five-point calibration curve at 50, 100, 200, 500 and 1000 ng/mL PFOS standards was constructed using the combustion method as samples and exhibited good linearity with R2>0.9999. Quantification was based on external calibration. The analytical conditions for ion chromatography have been reported elsewhere (Kärrman et al. et al., 2021).
Mass balance analysis approach
The measured PFOA concentration (ng/mL) in the sample extract of the EOF analysis were converted to the corresponding fluoride-equivalent concentration (ng F/mL) using the following formula:
CF_PFAS = CPFAS * NF * AF/MWPFAS (1)
where CPFAS is the concentration of the target compound (i.e., PFOA), NF is the number of fluorine atoms in the target compound (i.e., 15), AF is the atomic weight of fluorine (g/mol, i.e., 19) and MWPFAS is the molecular weight of the target compound (i.e., 413). The sum of known extractable fluorine concentration (ΣCF_PFAS) was calculated by summing the fluorine concentrations from all individual PFASs. Values below limits of quantification (LOQ) were set for calculating ΣCF_PFAS. Levels of unidentified organic fluorine were calculated by subtracting EOF from all quantifiable PFAS, which in this present study only PFOA.
Quality assurance and quality control (QA/QC)
In PFOA analysis, the method detection limits (MDLs) were determined as three times the signal in the negative control; and in absence of the analyte in the blank the lowest point in the calibration curve, which ranged 0.02-0.04 ng/mL for most of the PFOA. To ensure stable sensitivity over the entire instrumental analysis a quality assurance (QA) sample made of PFOA standards (2 ng/mL) was injected between each eight samples; the relative standard deviation of the intensity of QA samples was found to be below 10%. Before real sample analysis, matrix spike recoveries were conducted by spiking 1 ng of PFOA into different tissues (e.g. plasma, liver, lung, lung lavage and kidney) of the control subjects, the accuracy of the method was evaluated by subtracting the spiked level from the non-spiked sample and then divided by the spiked level times 100%; results were found to be between 92 to 111% (Table S1, Supplementary Information).
For the analysis of EOF, multiple measurements of combustion blanks were effectuated and repeated until the combustion blanks showed low variability (below 5% relative standard deviation) over the last three combustion blanks, so as to reduce the CIC system contained background fluoride contamination. All measurements of samples were first subtracted from the combustion blanks between samples before quantification, using an external calibration curve. An instrumental standard (PFOS 480 ng F/mL) was analyzed to evaluate the whole performance of the CIC. Signal fluctuation (RSD: 15%) was observed in the instrument standard in every five samples. A spiked plasma sample containing 200 ng F/mL was extracted in triplicate and the recovery were found to be 80 ± 9% and the relative standard deviation was found to be 8%.