Derivatization‐free determination of short‐chain volatile amines in human plasma and urine by headspace gas chromatography‐mass spectrometry

Abstract Background Short‐chain volatile amines (SCVA) are an interesting compound class playing crucial roles in physiological and toxicological human settings. Dimethylamine (DMA), trimethylamine (TMA), diethylamine (DEA), and triethylamine (TEA) were investigated in detail. Methods Headspace gas chromatography coupled to mass spectrometry (HS‐GC‐MS) was used for the simultaneous qualitative and quantitative determination of four SCVA in different human body fluids. Four hundred microliters of Li‐heparin plasma and urine were analyzed after liberation of volatile amines under heated conditions in an aqueous alkaline and saline environment. Target analytes were separated on a volatile amine column and detected on a Thermo DSQ II mass spectrometer scheduled in single ion monitoring mode. Results Chromatographic separation of selected SCVA was done within 7.5 minutes. The method was developed and validated with respect to accuracy, precision, recovery and stability. Accuracy and precision criteria were below 12% for all target analytes at low and high levels. The selected extraction procedure provided recoveries of more than 92% from both matrices for TMA, DEA and TEA. The recovery of DMA from Li‐heparin plasma was lower but still in the acceptable range (>75%). The newly validated method was successfully applied to plasma and urine samples from healthy volunteers. Detected concentrations of endogenous metabolites DMA and TMA are comparable to already known reference ranges. Conclusion Herein, we describe the successful development and validation of a reliable and broadly applicable HS‐GC‐MS procedure for the simultaneous and quantitative determination of SCVA in human plasma and urine without relying on derivatization chemistry.


| INTRODUC TI ON
Dimethylamine (DMA), trimethylamine (TMA), diethylamine (DEA) and triethylamine (TEA) are short-chain aliphatic amines (SCVA) and used as biomarkers for the identification of different physiological, pathophysiological and toxicological states in human. DMA and TMA are present in several human body fluids like urine, blood and sweat. 1 Bacterial metabolism results in the transformation of phosphocholine (PC) and choline-derived from meat, seafood, and dairy products-to TMA. [2][3][4] In recent years, the non-odorous and non-volatile oxidation product of TMA-trimethylamine-N-oxide (TMAO)-attracted the attention of physicians and clinical chemists worldwide. [5][6][7][8] TMAO is suspected to be of diagnostic or prognostic value, respectively, in cardio vascular events and pneumological diseases like community acquired pneumonia. 9,10 Oxidation of TMA to TMAO is catalyzed by flavin-monooxygenases (FMO) in liver microsomes. 11 Genetic mutations and liver damage can lead to a loss in FMO activity and result in trimethylaminuria, also known as fish odor syndrome. Patients suffering from this condition are characterized by penetrant smell of fish resulting from the transpiration and expiration of excess TMA. Until now, no cure from this disease is reported and the only way to ameliorate the symptoms is to avoid choline-rich nutrition. 12 DMA is present in a variety of nutrients. An increase of the urinary DMA concentration can be observed, for example, after the consumption of seafood. 13 In humans, degradation of glycine and sarcosine and subsequent methylation of monomethylamine lead to DMA formation. Additionally, DMA is a metabolite of asymmetric dimethyl arginine (ADMA). Asymmetric dimethyl arginine is known as inhibitor of nitrogen oxide synthase, playing a significant role in several pathophysiological states like renal diseases and chronic obstructive pulmonary disease. 14,15 DEA and TEA are closely related short-chain amines which are not derived metabolically. However, DEA and TEA are widely distributed reagents in pharmaceutical and chemical industries. 16 TEA irritates human mucous membranes and causes headache and nausea. 17 Nitrosation of DMA and DEA results in the formation of N-dimethyl-nitrosamine (NDMA) and N-diethylnitrosamine (NDEA), respectively. Both compounds are classified as cancerogens. [18][19][20] NDMA and NDEA are currently found as impurities in several active pharmaceutical ingredients (eg, Valsartan and Irbesartan) and perturb national authorities. [21][22][23] Formation of nitrosamine is also reported in human and animal gastric juice and is accelerated by bacteria in urinary tract infections. 18,24 SCVA were mostly detected by gas chromatography (GC) after direct injection. 1 Additionally, determination of TMA is also performed by liquid chromatography. 25 However, in food and bioanalysis, headspace (HS)-GC is preferred due to (a) the low contamination of the GC column and injection system and (b) to the high sensitivity and repeatability. In the context of SCVA determination, different derivatization procedures were described for quantification. 26,27 Although chemical derivatization of target analytes is a common preanalytical procedure, this additional step is often time-consuming and causes an additional source of error. In combination with HS-GC, dynamic extraction techniques-like solid-phase microextraction (SPME)-are often reported to further improve detection limits of SCVA. However, SPME needs a higher grade of equipment. 28 In here, we report a derivatization-free and static HS-GC method using a volatile amine column for human plasma and urine samples, respectively. To the best of our knowledge, no HS-GC-MS assay for the simultaneous analysis of the selected SCVA has been reported in human body fluids. The method was validated according to international guidelines 29 and can easily be applied even in less equipped laboratories offering short preparation and analysis times.

| Materials
Analytical reference standard of TMA was obtained from Sigma-Aldrich. Dimethylamine (2 mol/L solution in methanol), DEA, TEA, isopropylamine, carnitine, and choline were also purchased from Sigma-Aldrich and were of highest analytical grade. Deuterated TMA d9 was obtained from Toronto Research Chemicals. Sodium hydroxide (NaOH), hydrochloric acid solution (HCl), and potassium chloride (KCl) were purchased from VWR. Phosphate-buffered saline (PBS) was from Gibco Life technologies. Headspace vials, aluminum caps, and MSsepta were purchased from Infochroma. Pure water was generated from an in-house water purification system from Labtec. For all experiments, Gilson pipettes and Gilson DIAMOND tips were used. Lithiumheparin tubes without gel separator were from BD (Becton Dickinson).

| Apparatus
All samples were analyzed using a Focus Trace GC Ultra with Triplus Headspace injection system and DSQ II MS detector (Thermo Scientific). Chromatographic separation was performed on a Restek Rtx-Volatile Amine column (30 m; 0.32 mm ID; 5 µm; BGB). Helium flow was set to constant flow at 2 mL/min. Split ratio was set to 7.
The starting temperature for the oven was 40°C and held for 4 minutes. Temperature was increased with 25°C/min to 250°C and kept constant for further 3 minutes. Headspace conditions were as follows: agitating for 10 minutes at 70°C. Syringe temperature was set to 80°C and 2 mL of gaseous sample were drawn from the head-

| Preparation of calibration and QC samples
Separate stock solutions for multi-analyte calibration (Cal) and quality control (QC) samples were prepared in 0.1 mol/L HCl. All solutions were stored in aliquots at −20°C. The final calibration concentrations are given in Table 1 for each target analyte. Isopropylamine (IPA) was used as internal standard (100 µmol/L in 0.1 mol/L HCl).

| Sample preparation
Cal, QC, and sample preparation was performed on ice. As negative control, Li-heparin tubes were filled with PBS (10 mL), vortexed for 10 seconds, and incubated for 10 minutes. For analysis, 400 µL, Cal, QC, or authentic samples were mixed with 10 µL internal standard mix and 750 µL 2 mol/L NaOH/0.5 mol/L KCl in a 20 mL GC headspace vial and sealed directly. Samples were vortexed for 5 seconds and set on the bench to reach room temperature. Afterward samples were analyzed as described above. Choline and carnitine solutions were prepared in PBS (5 mmol/L) and analyzed accordingly.

| Method development
In the following, QC Med was used for evaluating different HS conditions. Incubation time was investigated in 10 minutes steps from 10 to 30 minutes. Incubation temperature was tested at 60, 70, and 80°C. For liberation of the free amine, NaOH (0.5-2.0 mol/L) and KCl (0.3-0.5 mol/L) were used. Different sample volumes were evaluated (50 µL, 100 µL, 250 µL, and 400 µL).

| Data analysis
Thermo Excalibur software (2.2 SP1.48) was used for peak integration and quantification of data. GraphPad Prism 7 (GraphPad Software) was used for statistical analysis and illustrations.

| Method validation
Initially, TMA d9 was used as internal standard. Since we observed a high inter-assay variance of signal intensity, IPA replaced TMA d9 as internal standard for the method validation and all further assays.
Isopropylamine could clearly be separated from all other SCVA. The use of internal standards corrected for any apparent loss of analytes during the liberation, static headspace extraction, and split injection.
Calibration curves using six concentration levels with six replicates each were constructed to evaluate the calibration model. Calibration ranges for all analytes are given in Table 1. Accuracy was given in terms of bias as the percent of deviation of the mean calculated con-

| Applicability
The developed method was successfully applied for selected SCVA determination in human plasma and urine from 11 healthy volunteers. Samples of the same individual were analyzed in the same run.
As shown in Table 2 26 To exclude production process related contamination with SCVA, Li-heparin tubes were filled with 10 mL PBS as negative control. No contamination with SCVA was detected. We decided to use Li-heparin tubes without separator gel for blood sampling since gel separator tubes may affect analytical results. 39 As reported by Bain et al, 40 quaternary ammonia compounds can undergo a Hofmann-elimination resulting in ex vivo generation of TMA. Therefore, carnitine and choline were prepared in PBS (5 mmol/L each) and analyzed according to the described method.

| D ISCUSS I ON
Although the applied concentration is up to 100-fold higher as reported for plasma, 5 only a slight TMA signal was detected under current HS conditions ( Figure 4). One may speculate if incubation temperature of 50°C and missing salting out conditions-as reported by Bain and coworkers-are not sufficient to equilibrate the HS. Therefore, plasma TMA values, reported in here, are higher compared to those reported previously. 40 As expected, no DEA was found in urine or plasma, respectively. Interestingly, in only one urine sample TEA was calculated to 0.5 µmol/L. Identity was confirmed by retention time and fragmentation pattern ( Figure   S4). The occurrence of TEA will be inspected in detail but is not further discussed in here.
In conclusion, the presented analytical method allows the simultaneous accurate and precise quantification of four pharmacologically and toxicologically important short-chain volatile amines in Li-heparin plasma and urine. The lack of derivatization further accelerates the preanalytical phase and avoids errors in sample preparation, respectively.
The method met the validation criteria for all analytes. Subsequently, we demonstrated its practicability by the successful analysis of 11 paired Li-heparin plasma and urine samples of healthy volunteers. We aim to investigate SCVA in further clinical and toxicological studies.
The results of these studies will be presented elsewhere.