2.1 Chemicals and reagents
Kallidin trifluoroacetic acid (TFA) salt (96.9%, HPLC; Tocris, Bristol, UK), bradykinin acetate (99.0%, HPLC; Sigma-Aldrich, St. Louis, MO, USA), and their metabolites des-Arg(9)-bradykinin acetate (98.7%, HPLC; Santa Cruz Biotechnology, Dallas, TX, USA), bradykinin 1‑7 TFA salt (≥ 95.0%, HPLC; GenScript, Piscataway Township, NJ, USA), bradykinin 1-5 TFA salt (≥ 95.0%, HPLC; GenScript), bradykinin 2-9 TFA salt (≥ 95.0%, HPLC; GenScript), and des-Arg(10)-kallidin TFA salt (95.9%, HPLC) were used in this study. [Phe8Ψ(CH-NH)-Arg9]-bradykinin TFA salt (97.5%, HPLC) was applied as the internal standard. Formic acid (FA, ≥ 98%) and TFA (100.3%) were supplied by Sigma Aldrich. HPLC-grade methanol, water, and dimethyl sulfoxide (DMSO, ≥ 99.9%), and MS-grade methanol and ammonium acetate (99.5%) were obtained from Fisher Scientific (Loughborough, UK). Furthermore, HPLC-grade acetonitrile (ACN, Applichem, Darmstadt, Germany), MS-grade water (Honeywell Fluka, Seelze, Germany), and ammonia (30.9%) (VWR Chemicals, Radnor, PA, USA) were utilized. Isotonic saline solution 0.9% was provided by B. Braun (Melsungen, Germany).
Sampling of NLF in healthy volunteers was performed in compliance with the ethical principles of the Declaration of Helsinki and was approved by the ethics committee of the medical faculty at the Heinrich-Heine-University Duesseldorf (study number: 6112). Written informed consent was obtained from all participants before enrolment.
2.2 Preparation of stock and working solutions
Lyophilized kinin peptides were dissolved and diluted separately in 0.3% TFA in 25/75 ACN/water (v/v/v) prior to the preparation of a combined working solution containing 500 ng/mL of each peptide salt. [Phe8Ψ(CH-NH)-Arg9]-bradykinin as an internal standard was dissolved in 0.1% FA in water (v/v) and subsequently diluted to achieve a working solution of 500 ng/mL in 0.3% TFA in 25/75 ACN/water(v/v/v). All peptide solutions were prepared using low protein-binding tubes (Sarstedt, Nümbrecht, Germany).
2.3 Sample preparation
A 0.9% isotonic saline solution was used as blank surrogate matrix for the respiratory saline lavage fluids. Owing to the endogenous presence of kinins and the long half-life of bradykinin 1-5, no reliable kinin-free human blank matrix could be generated. Optimized inhibitors were applied to effectively prevent the generation and degradation of the kinin peptides, based on previously published suitable inhibitor cocktails.27 SPE was performed by applying 96-well Oasis weak cation exchange (WCX) µ-elution plates (Waters, Milford, MA, USA). All cartridges were conditioned with 200 µL of methanol, followed by 200 µL of 5% aqueous ammonium hydroxide (v/v). Subsequently, the wells were prefilled with 150 µL of 3 ng/mL internal standard in 5% aqueous ammonium hydroxide (v/v), and 100 µL of sample was then loaded. Washing was performed with 300 µL of 5% aqueous ammonium hydroxide (v/v) and 300 µL of 10% methanol in water (v/v). Elution was conducted three times with 50 µL of 1% TFA in 75/25 ACN/water (v/v/v). The resulting eluates were evaporated to dryness under a gentle stream of nitrogen at 60 °C while shaking at 300 rpm. The residues were dissolved in 75 µL of 10/10/80 FA/methanol/water (v/v/v).
Chromatography was performed on an Agilent 1200 SL series system (Agilent Technologies, Ratingen, Germany) consisting of a degasser (G1379B), a binary pump SL (G1379B) and a column oven TCC SL (G1316B). A Phenomenex SynergiTM 2.5 µm Hydro-RP 100 Å column (100x2.0 mm; Torrance, CA, USA) was used for the chromatographic separation. The mobile phases were composed of water and methanol (B) both containing 3.2% DMSO and 0.1% FA (v/v). A 7.5 min binary gradient was applied, maintaining the amount of mobile phase B at 5% for 1.5 min, increasing it to 20% until 2.2 min, to 27% until 2.7 min, to 35% until 3.1 min, and finally to 95% after 6.2 min. Mobile phase B was kept constant at 95% until 6.7 min before decreasing it to 5% and column re-equilibration for 3 min. The flow rate was set to 400 µL/min, and the injection volume of 50 µL was applied with an HTC PAL autosampler (CTC Analytics AG, Zwingen, Switzerland). After every injection, the autosampler syringe was rinsed twice with 0.2% FA in 80/20 ACN/water (v/v/v). Samples were stored at 18 °C in the autosampler.
The LC system was coupled to an API 4000 (AB Sciex, Darmstadt, Germany) mass spectrometer equipped with a Turbo V source for detection. The electrospray interface wasoperated in positive mode with multiple reaction monitoring mode. The curtain gas was maintained at 31 psi, the collision gas at 8 psi, the nebulizer gas at 45 psi, the heater gas at 65 psi, the ion spray voltage at 5500 V, and the source temperature at 350 °C. Peptide-specific parameters are displayed in table 1.
Data acquisition was conducted using Analyst® 1.6.2 software (AB Sciex, Darmstadt, Germany), and raw data evaluation was performed using MultiquantTM 3.0.2 (AB Sciex, Darmstadt, Germany).
2.5 Method development
2.5.1 Adaption of optimized injection solvent and sample collection material
A previously conducted DoE approach to optimize the injection solvent conjointly with the sample collection material to reduce non-specific adsorption of bradykinin and thus increase sensitivity,26 had to be adapted to avoid peak broadening or breakthrough of the more hydrophilic kinin peptides. By using the D-optimal optimization model, an amount of 5–20% organic fraction in the injection solvent was investigated. This range correlated with the binary gradient, as no breakthrough of the kinins was expected. Furthermore, the calculations included a maximum intensity loss of 15% of the predicted intensity of the optimized injection solvent, based on current bioanalytical guidelines for the accuracy limits.28 Six injection solvent compositions were calculated with distinct organic fractions and analyzed in triplicates by LC-MS/MS measurement, and responses and peak shapes were compared to the original injection solvent for bradykinin only (8.7% FA in 5.3/36.6/49.4 methanol/DMSO/water (v/v/v/v)).
2.5.2 Improvement of SPE recovery
The method development focused on maximizing recovery to reduce peptide loss during washing steps and to enable the detection of endogenous concentrations in the low pg/mL range. A previously developed SPE protocol for bradykinin only26 had to be adapted, since the kinin peptides differ in their amounts of hydrophobic and positively charged amino acids (fig. 1). Mixed-mode strong cation exchange (MCX) and WCX µelution SPE were considered to evaluate which material would fit best to all analytes. All experiments evaluating the washing and elution solvents were conducted using neat solution in duplicate.
Bioanalytical method validation was conducted considering the regulatory bioanalytical guidelines of the US Food and Drug administration.28 Linearity, accuracy, precision, sensitivity, recovery, matrix effect, carry-over, and stability were considered during the validation process.
Linearity was determined in six runs using nine to eleven distinct calibrator levels (depending on the lower limit of quantification (LLOQ) of the individual peptide), which were analyzed in single determinations. In compliance with bio-analytical guidelines, the actual concentration of 75% of all calibration curve standards had to deviate less than ±15% (±20% at the LLOQ) from their nominal concentration (relative error (RE)).28
2.6.2 Accuracy, precision and sensitivity
Accuracy and precision were assessed using up to seven quality control (QC) levels covering the whole calibration range on three distinct days. The number of QCs depended on the magnitude of the calibration range per peptide. Five replicates per QC level were analyzed each day. Accuracies were determined as the deviation of the actual concentration from the nominal concentration (RE) for within-run and for between-run accuracy. Using one-way ANOVA, within-run precision was calculated as repeatability and between-run precision as day-different intermediate precision (coefficient of variation (CV)). In line with regulatory guidelines, accuracy and precision were not allowed to exceed ±15% (±20% at the LLOQ).28 The signal-to-noise ratio (S/N) had to be higher than 5:1 at the LLOQ. The limit of detection (LoD) was further calculated as follows using the results from six calibration curves (equation (1):29
(1) Calculation of limit of detection. σ: standard deviation of the y-intercept, S: mean slope of the calibration curves
Carry-over was evaluated by alternatingly injecting blank samples and upper limit of quantification (ULOQ) calibration curve standards six times. According to regulatory bioanalytical guidelines, carry-over in the blank sample following the ULOQ was not allowed to exceed 20% of the LLOQ signal and 5% of the internal standard signal.28
2.6.4 Recovery and absolute matrix effect
Recovery was determined at four distinct QC levels covering the calibration curve range (high, middle, low and around the LLOQ) in triplicate. Pre-spiked extracted samples were compared to processed blank samples spiked with the same concentrations after µelution SPE. Matrix effects of the peptides were analyzed at the same four distinct QC levels by comparison of post-spiked samples to neat solutions (n=3).
Stability studies were conducted at four QC levels (high, middle, low and around the LLOQ) under different storage conditions. Benchtop stability was investigated by placing the prepared QC samples at room temperature for one and three hours (n=5). Freeze-thaw stability (room temperature to -80 °C) was evaluated by analyzing the QC levels after one and three cycles (n=5). Between each cycle, samples were frozen for at least 12 hours. The autosampler stability was assessed by keeping QC samples in the autosampler at 18 °C for 18 hours and then repeating the QC sample measurement. Finally, short-term stability of processed and evaporated samples was analyzed after storage for 24 hours at +4 °C (n=5). Stability at the specific conditions was proven if the mean concentration at each level did not exceed ±15%. To evaluate the stability of the analyte working solution, peak areas of the kinin peptides on 15 distinct days after 15 freeze-thaw cycles, measured routinely during method performance qualification (1 ng/mL, neat solution), were analyzed for their CV (acceptance criterion ≤ 15%).
Nasal lavage with 10 mL of 0.9% isotonic saline (5 mL per nostril) was performed in a healthy female volunteer. The volunteer was asked to tip her head backwards, hold the breath, and refrain from swallowing. The obtained fluid was collected directly into the inhibitor cocktail and was immediately vortexed after completing the sampling. The samples were centrifuged at 4 °C for 15 min at 500 ×g to remove cells, mucus, and debris. A 100 µL aliquot of the supernatant was analyzed.