Automated extraction of PFAS from human serum using Biotage® Extrahera™ LV-200 and ISOLUTE® PLD+ for PFAS

For research use only. NOT for use in diagnostic procedures.

Per and polyfluoroalkyl substances (PFAS) comprise many compounds that occur in a broad range of matrices and environments. PFAS are of concern because of their high persistence, bioaccumulation and slow elimination, and impacts on human and environmental health. Exposure to PFAS correlates with changes in metabolism, higher cholesterol, and increased risk of some cancers. PFAS pose challenges in the analytical laboratory as they are present in common consumables and hardware. The method described in this application note can be used to reproducibly quantitate multiple PFAS classes with low matrix effects at clinically relevant levels in human serum.
Figure 1. Selected PFAS structures

Introduction

This application note describes the extraction of thirty-one PFAS from human serum using Biotage® ISOLUTE® PLD+ for PFAS plates prior to UHPLC-MS/MS analysis.

The sample preparation procedure simultaneously removes proteins and phospholipids from serum, maintaining high, reproducible analyte recoveries, and minimizing matrix interferences. ISOLUTE® PLD+ for PFAS can be integrated quickly and easily into routine workflows, increasing productivity and reducing instrument downtime. No evaporation step is needed, extracts are diluted with compatible solvent prior to injection.

The application note includes optimized conditions for automated processing of ISOLUTE® PLD+ for PFAS plates (using the Biotage® Extrahera™ LV-200, see appendix for settings) and manual processing (using the Biotage® VacMaster™ -96 vacuum manifold). Data generated using both processing systems is shown.

Analytes

A wide range of PFAS classes were selected to reflect the broad applicability of this simple sample preparation approach.

Table 1. Target PFAS analytes

Sample preparation procedure

Format

ISOLUTE® PLD+ for PFAS Plate, part number 919-0050-P01

Processing

Samples were processed using a Biotage® Extrahera™ LV-200 automated sample preparation workstation, or manually using a Biotage® VacMaster™ 96 sample processing manifold.

Note: ISOLUTE® PLD+ for PFAS plates can also be processed using the Biotage® Pressure+96 Positive Pressure Manifold. Processing parameters are available on request.

Sample pre-treatment

Approximately 150 µL of pooled serum (sample) was added to each well of a 2 mL square collection plate for automated or manual processing. If used, add internal standard at this stage. For example, an appropriate concentration in 10 µL methanol or acetonitrile.

Extraction procedure

Extraction is performed using ISOLUTE® PLD+ for PFAS using a 7:1 (v/v) solvent:sample ratio with ‘solvent first’ methodology.

Biotage®  Extrahera™  LV-200:

Dispense 700 μL of acetonitrile (MeCN) extraction solvent in each well. Dispense 100 µL of serum sample into each well. Solvent/sample will not flow until pressure is applied, allowing for efficient in-well mixing. Mix the solvent and sample with 2x aspirate/dispense cycles and wait 5 minutes. Process the plate by applying 0.4 Bar (5 min), 0.8 Bar (1 min). Collect the extracts in a 2 mL square collection plate. See Appendix 1 for additional details.

Biotage®  VacMaster™-96:

Using a multi-channel pipette (or similar), dispense 700 μL of acetonitrile (MeCN) extraction solvent in each well. Dispense 100 µL of serum sample vertically into each well with force. Solvent/sample will not flow until vacuum is applied, allowing for efficient in-well mixing. Mix the solvent and sample with 5x aspirate/dispense cycles and wait 5 minutes. Process the plate by applying -0.2 Bar (5 min), -0.4 Bar (2 min). Collect the extracts in a 2 mL square collection plate.

Note: This process can be scaled down to for a reduced serum sample volume of 50 µL. 350 µL of acetonitrile should be used as extraction solvent, following the same procedure. A summary of method performance for 50 µL samples can be found in table 4 (automated method) and table 5 (manual method).

Post extraction

Using a multi-channel pipette (or similar), dispense 200 µL of extracted sample to each well of a new 2 mL square collection plate. Dilute the contents of each well with 200 μL of 20 mM ammonium acetate (aq) dilution solvent. Vortex the plate gently for 30 s to mix the contents and cover with a pierceable sealing cap before transferring to the LC-MS/MS system for analysis.

Analytical conditions

U/HPLC parameters

• Instrument: Shimadzu Nexera UHPLC using fluoropolymer-free tubing and a pre-injector trap column (Restek PFAS Delay 50 x 2.1 mm).
• Column: Avantor ACE Ultracore SuperC18 2.5 µm (50 x 2.1 mm)
• Raptor ARC-18 EXP guard 2.7 µm 5 x 2.1 mm
• Mobile phase A: 5 mM ammonium acetate (aq)
• Mobile phase B: 5 mM ammonium acetate in MeOH
• Flow rate: 0.4 mL/min
• Column temperature: 40 °C
• Injection volume: 5 µL (no rinsing)
• Sample temperature: 15 °C

MS/MS parameters

• Instrument: AB Sciex 5500 triple quadrupole system operating in negative ion mode
• Ionspray voltage: -3500 V
• Source temperature: 500 °C
• Curtain gas: 40 psi
• Source gases: GS1 40 psi / GS2 60 psi

Table 2. Gradient parameters

Table 3. MRM parameters

Results

This application note was developed using pooled human serum. Further optimisation may be required for other similar matrices (for example, to compensate for regional dietary variation). Recovery data shown in this application note was generated using intact, non-stripped, representative matrix.

Extraction recovery was determined using a 160 pg (1.6 ng/mL extracted) spike before extraction as a proportion of the spike after extraction (fort). The spike area response was used to determine extraction repeatability as % RSD (n=6). Matrix effects were estimated for each analyte using the fort as a proportion of a dilute standard at the same concentration (0.5 pg on-column). Blank contributions were estimated for each analyte using the blank response (n=3) as a proportion of the dilute standard (0.5 pg on-column).

Recovery and reproducibility: automated method

Typical analyte recovery processed using Extrahera™ LV-200 was between 80% and 94% for a 100 µL sample load. Extraction repeatability with this system is typically less than 5% RSD (n=6) (figure 2).
Figure 2. Recovery (%RSD (n=6) shown as error bars) of PFAS analytes from 100 µL serum (spiked at a concentration of 1.6 ng/mL), using the method described in this application note.

Typical matrix factors for samples processed using Extrahera™ LV-200 were between 1.0 and 1.5 (figure 3). Matrix effect factors for PFOA, PFHxS, and PFOS are elevated above 1.5 due to contribution of these PFAS species present in the blank (non-stripped) serum.
Figure 3. Typical matrix factors for 100 µL serum processed using Biotage® Extrahera™ LV-200 relative to 1.6 ng/mL. This includes contribution from the sample matrix (pooled human serum, non-stripped), plus any from processing using Extrahera™ LV-200, ISOLUTE® PLD+ for PFAS plate and reagents.

Matrix blank factors for PFOA, PFHxS, and PFOS were between 0.4 and 1.0 (shown in figure 4). This matrix-derived contribution corresponds to between 0.6 and 1.6 ng/mL extracted sample (analyte dependent).

However, processing residues using ISOLUTE® PLD+ for PFAS are extremely low, as demonstrated by figure 5. This shows the combined contribution of PFAS species derived from the ISOLUTE® PLD+ for PFAS, processing using Extrahera™ LV-200, and the reagents used as described in this application note (with no matrix present), estimated as equivalent to ~22 pg/mL sample.
Figure 4. Matrix-derived blank contribution corresponds to between 0.6 and 1.6 ng/mL extracted sample (figure 4). Note: matrix consists of pooled human serum, suggesting that PFAS residues of up to 1.6 ng/mL (analyte dependant) are present in the sample matrix.
Figure 5. Contribution to the blank of PFAS species related to the ISOLUTE® PLD+ for PFAS extraction product, automated processing and reagents– with no matrix present. A process factor of 0.014 was determined, equivalent to ~22 pg/mL.

Recovery and reproducibility: manual method

Typical analyte recovery processing manually using vacuum was between between 80% and 93% for a 100 µL sample load (data not shown). Manually processed extraction repeatability was typically less than 5% RSD (n=6). Matrix factors were typically between 1.0 and 1.5. Matrix factors for PFOA, PFHxS, and PFOS are elevated above 1.5 due to contribution from the blank. Blank factors for PFOA, PFHxS, and PFOS were between 0.4 and 0.83, equivalent to between 0.6 and 1.3 ng/mL extracted sample. Process factors are similar to automated methods.

Automated vs Manual processing

PFAS Recovery and matrix factors when extracted using ISOLUTE® PLD+ for PFAS are comparable when comparing automated processing using the Extrahera™ LV-200 to manual processing using a VacMaster™-96. PFAS repeatability is comparable when using 100 µL sample load; automated processing using the Extrahera™ LV-200 demonstrates improved repeatability compared to manual processing for lower sample load volumes.

Linearity and LOQ

Method performance was evaluated using external standards extracted from spiked matrix over 8 levels from 0.1 ng/mL to 100 ng/mL (n=4). Limit of quantitation (LOQ) was estimated as the lowest extracted concentration with signal/noise (S/N) > 10-20. Analyte linearity was determined acceptable where: the calibration coefficient (r2) was > 0.995; S/N > 10-20 (estimated using Analyst 1.6.3, peak-to-peak); repeatability < 10% RSD (< 15% at LOQ), with accuracy 90-110% (80-120% at LOQ).

A representative chromatogram of matrix extracted standards spiked at 2 ng/mL is shown in figure 5. All analytes demonstrate good separation and peak shape, the additional peak in the MRM at 1.9 minutes is probably due to a branched isomer of PFPeA present in the serum matrix.
Figure 6. Representative PFAS chromatogram, extracted serum, 2 ng/mL spike

Representative calibration curves are shown below (figures 7 & 8, automated and manual processing respectively). Method performance for all analytes is tabulated below (table 4 & 5, automated and manual processing respectively). Most analytes demonstrate LOQ at 0.1 ng/mL. However, some perfluorocarboxylic acid and perfluorosulfonic acid LOQ are
0.4 ng/mL. All analytes demonstrate good linearity, r2 > 0.997. The majority of analytes demonstrate repeatability < 10% at all calibration levels (< 20% at LOQ). Typical analyte accuracy was 90-110% (80-120% at LOQ).
Figure 7. Calibration curves for a) PFOA b) PFOS c) PFOSA and d) GenX extracted from 100 µL of serum using Biotage® Extrahera™ LV-200
Figure 8. Calibration curves for a) PFOA b) PFOS c) PFOSA and d) GenX extracted from 100 µL of serum using Biotage® VacMaster™ -96

Table 4. Method performance for the automated method (100 µL and 50 µL sample volumes)

Table 5. Method performance for the manual method (100 µL and 50 µL sample volumes)

Extract cleanliness

Extract cleanliness (matrix depletion) is demonstrated in figure 9. 100 µL serum was precipitated 1:7 (v/v) with ACN and either centrifuged or extracted using ISOLUTE® PLD+ for PFAS. The supernatant or extract was dried and reconstituted in an equal volume of 30% MeOH prior to injection onto the LC-MS/MS system. A combined phospholipid (PL) and lysophospholipid (LPL) profile was generated from the TICs of phosphatidylcho- line and lysophosphatidylcholine MRM transitions respectively. ISOLUTE® PLD+ for PFAS extraction is an extremely efficient means of phospholipid depletion, resulting in 99.9% depletion of LPL and PL compared to precipitation and dilute/shoot.
Figure 9. Full range overlaid MRM with 1.5E7>1.5E5 expanded inset, for serum prepared by precipitation with acetonitrile (1:7, v/v)(blue), compared with serum prepared using ISOLUTE® PLD+ for PFAS (red)

Discussion & Conclusion

Using ISOLUTE® PLD+ for PFAS for sample preparation, this application note demonstrates high PFAS recovery and sensi- tivity with low matrix factors and good repeatability.

The use of a ‘crash and filter’ flow-through strategy incorpo- rating a multifunctional sorbent that provides simultaneous removal of >99.9% proteins and phospholipids has several advantages compared to alterative solid phase extraction- based processes.

Serum extracted using ISOLUTE® PLD+ for PFAS results in extremely clean extracts, prolonging LC column lifespan and minimising instrument down time for cleaning and maintenance.

The simplicity of the method, with fewer extraction steps, and no evaporation requirement has additional benefits. Throughput is increased, with up to 96 samples ready for analysis is ~35 minutes when processed using Extrahera™ LV-200, compared with ~ 80 minutes for a WAX based SPE approach (includes 15 mins evaporation time). This, combined with reduced reagent preparation time, has significant produc- tivity benefits.

The range of PFAS species that can be effectively extracted using the ISOLUTE® PLD+ for PFAS approach is extended. Elimination of the evaporation step means that losses of sulfonamide and subtituted sulfonamide PFAS, which can demonstrate low or no recovery following evaporation and reconstitution, are reduced. Poor recoveries of longer chain PFAS associated with WAX based catch and release SPE are also eliminated.

The ubiquitous nature of PFAS meant that the sample matrix (non-stripped pooled human serum) used to generate data in this application note was found to contain up to 1.6 ng/mL of some PFAS species. However, we found that the ISOLUTE® PLD+ for PFAS plate, processing system and reagents typically contributed ~22 pg/mL, making it suitable for determination of PFAS in serum samples at clinically relevant levels.

Chemicals and reagents

• PFAS stock standards were purchased from Wellington Laboratories (Guelph, Canada). The suite contains 10 classes of PFAS, varying by functionality, including: carboxylic acids, sulfonic acids and telomers, sulfonamides, and ethoxy compounds.
• Mixed intermediate standards were prepared from the individual stocks on a class by class basis e.g. carboxylic acids. The intermediate standards were prepared in the appropriate solvent indicated on the stock certificate of analysis (e.g. carboxylic acids require 0.4 mole equivalent of base to prevent formation of methyl esters). Some stocks are sold as potassium or sodium salts e.g. sulfonic acids. A correction should be applied to the dilution to calculate the final concentration of free acid in the mixed intermediate. The intermediate stocks were prepared at 4 µg/mL and stored at -20 °C in 1.5 mL polypropylene vials.
• A mixed working standard was prepared weekly at 160 ng/mL in MeOH from the mixed intermediates and stored at -20 °C in 1.5 mL polypropylene vials. As a guide, a 1/25 dilution can be prepared using: 40 µL x 4 µg/mL / 1000 µL = 160 ng/mL.
• Reagents were purchased from Merck Life Science (Gillingham, UK). LC/MS grade solvents were from Rathburn Chemicals Ltd. (Walkerburn, UK).
• Water (18.2 MΩ.cm) was drawn fresh daily from a Direct-Q 5 water purifier (Merck Life Science UK, Watford, UK).
• All reagents were prepared fresh daily in 250 mL polypropylene bottles.
• Ammonium acetate containing reagents were prepared from a 5 M NH4OAC stock (Sigma-Aldrich ammonium acetate solution, 09691-100ML) stored at 2-8 °C.
• 20 mM NH4OAC was prepared by diluting 1 mL 5 M stock in 250 mL freshly drawn water.
• Mobile phases A and B were prepared by diluting 0.5 mL 5 M NH4OAC in 500 mL water or MeOH respectively.
• Serum samples were pooled discard serum, obtained from Welsh Blood Service (Llantrisant, UK).

Additional information

• Additional precautions were taken to minimize the possibility of system contamination from lab and environmentally derived PFAS. Care should be taken with all consumables, in-house studies have demonstrated consumables manufactured from similar materials can have differing PFAS residue profiles.
• Contact with glass surfaces was minimized as PFAS bind to glass and may be released or transferred during subsequent processing. Consumables used were restricted to virgin polypropylene (PP) or high-density polyethylene (HDPE) where possible. Disposable nitrile gloves were worn at all times and changed frequently.
• Additional triplicate blank analyses were acquired during a typical experiment to determine potential sources of interference: a) solvent blank, freshly prepared solvents in the final 2 mL collection plate in the same proportions as the final dilution (ACN 1:1 20 mM NH4OAC); b) consumable blank, extraction media processed using freshly drawn water as the sample in place of matrix; c) matrix blank, unspiked sample matrix.
• Extrahera™ LV-200 solvent inlets were flushed with 4x 20 mL MeOH followed by 4x 20 mL ACN prior to use.
• The UHPLC inlet was modified for the determination of PFAS (see opposite, figure 10). FEP tubing was replaced with PEEK tubing where possible. PTFE-containing mobile phase filters were replaced with stainless steel filters. The vacuum degassers were bypassed. A PFAS delay column was installed between the pumps and injector. The inlet was flushed with methanol at 0.4 mL/min and 50 °C for 4 hours prior to use. The injector was flushed with methanol prior to use. Mobile phase and wash solutions are replaced daily.

Optional evaporation/reconstitution conditions

For increased analytical sensitivity the extract can be evaporated and reconstituted in a lower volume of solvent, suitable for injection to the analytical system. The following conditions are suggested:

Following elution, to each well, add 10 μL of keeper solvent (we recommend DMF, but DMSO or ethylene glycol may also be appropriate).

Transfer the collection plate to a TurboVap®-96 Dual evaporation system, and evaporate to constant volume (approximately 10 µL) at 40 oC using air or nitrogen at a flow rate of 36-60 L/min and a plate height 50 mm or greater.

Reconstitute in 10 mM ammonium acetate/methanol (1:1, v/v, 400 µL)
Figure 10. Modification to the UHPLC inlet, used for PFAS analysis.

Ordering information

Appendix: Biotage® Extrahera™ LV-200 method

The method described in this application note was automated on the Biotage® Extrahera™ LV-200. This appendix contains the software settings required to configure Extrahera™ LV-200 to run this method for a 100 µL sample. Details for 50 µL sample are avail- able on request. As described in the main body of the application note, analyte recoveries, linearities and LOQs were comparable for both manually processed (VacMaster™-96) and automated methods. Reproducibility was improved for samples extracted using the automated Extrahera™ LV-200 system. Total time for extraction of 96 samples using this method was 35 minutes.

                                 Screenshot                                          Settings

Literature Number: AN991