Abstract
Objectives
In this study we describe the development and validation of a liquid chromatography mass spectrometry method (LC-MS/MS) to quantify five tryptophan (TRP) metabolites within the kynurenine– and serotonin pathway and apply the method to serum samples of women in the first trimester of pregnancy. A secondary aim was to investigate the correlation between body mass index (BMI) and the five analytes.
Methods
A LC-MS/MS was developed for the analysis of TRP, kynurenine (KYN), 5-hydroxytryptophan (5-HTP), hydroxytryptamine (5-HT), and 5-hydroxyindole acetic acid (5-HIAA). Serum samples (n=374) were analyzed of pregnant women (median gestational age: 8 ± 2 weeks) participating in a subcohort of the Rotterdam Periconceptional Cohort (Predict study).
Results
The LC-MS/MS method provided satisfactory separation of the five analytes (7 min run). For all analytes R2 was >0.995. Within- and between-run accuracies were 72–97% and 79–104%, and the precisions were all <15% except for the between-run precisions of the low QC-samples of 5-HTP and 5-HT (both 16%). Analyte concentrations were determined in serum samples of pregnant women (median (IQR)); TRP (µmol/L): 57.5 (13.4), KYN (µmol/L): 1.4 (0.4), 5-HTP (nmol/L): 4.1 (1.2), 5-HT (nmol/L): 615 (323.1), and 5-HIAA (nmol/L): 39.9 (17.0). BMI was negatively correlated with TRP, 5-HTP, and 5-HIAA (TRP: r=−0.18, p<0.001; 5-HTP: r=−0.13, p=0.02; natural log of 5-HIAA: r=−0.11, p=0.04), and positively with KYN (r=0.11, p=0.04).
Conclusions
The LC-MS/MS method is able to accurately quantify kynurenine– and serotonin pathway metabolites in pregnant women, providing an opportunity to investigate the role of the TRP metabolism in the (patho)physiology of pregnancy.
Introduction
Tryptophan (TRP) is an essential amino acid found in dietary proteins, of which most is metabolized along the kynurenine– and serotonin pathway [1]. The kynurenine pathway is regulated by the hepatic tryptophan 2,3-dioxygenase (TDO) and the extra-hepatic indole amine 2,3-dioxygenase (IDO), which convert TRP into kynurenine (KYN). Tryptophan hydroxylase is the rate-limiting enzyme of the serotonin pathway converting TRP into 5-hydroxytryptophan (5-HTP), leading to the downstream metabolites 5-hydroxytryptamine (serotonin (5-HT)), 5-hydroxyindole acetaldehyde, and 5-hydroxyindole acetic acid (5-HIAA) (Figure 1) [2].
Kynurenine– and serotonin pathway of the tryptophan metabolism. The analytes are colored gray.
The kynurenine– and serotonin pathway are involved in many vital processes, such as immune response, neuroprotection, and regulation of vascular tone, which are important during pregnancy [3], [4], [5], [6]. Indeed, variations of kynurenine– and serotonin pathway metabolites have been associated with pregnancy complications, such as preeclampsia [7, 8]. Thus, simultaneous quantification of kynurenine– and serotonin pathway metabolites is highly relevant to gain more insight into the pathological mechanisms underlying these pregnancy complications, and to develop novel diagnostic and therapeutic targets for obstetrical care.
In recent years, a few methods have been described to determine kynurenine– and serotonin pathway metabolites in human serum using liquid chromatography (tandem) mass spectrometry (LC-MS/MS) [9], [10], [11], [12], [13], [14]. However, most of these methods did not include the simultaneous determination of TRP metabolites within the kynurenine– and serotonin pathway, and none of them applied their method in a large cohort of pregnant women. Here, we describe the development and validation of a LC-MS/MS method to quantify TRP metabolites, specifically TRP, KYN, 5-HTP, 5-HT, and 5-HIAA, and apply the method to serum samples of women in the first trimester of pregnancy. As a secondary aim we investigated the correlation between body mass index (BMI) and the five analytes, since it is known that BMI affects the TRP metabolism [15].
Materials and methods
Chemicals and reagents
The standards and internal standards TRP, KYN, 5-HTP, 5-HT, 5-HIAA, and 5-HT-d4 were purchased from Sigma-Aldrich (Saint Louis, Missouri, USA). TRP-d5 and 5-HIAA-13C6 were from Cambridge Isotope Laboratories (Tewksbury, Massachusetts, USA), 5-HTP-d4 from Toronto Research Chemicals (Toronto, Ontario, Canada), and KYN-d4 from Buchem B.V. (Apeldoorn, the Netherlands). LC-MS grade acetonitrile, formic acid, and methanol were purchased from Biosolve B.V. (Valkenswaard, the Netherlands), and the ultra-pure water was delivered in house by the Q-POD Ultrapure Remote Dispenser (Merck Millipore).
Standard solutions
The stock standard and internal standard solutions were prepared in methanol (4 mmol/L) with exception of KYN and KYN-d4, which were prepared in water. For internal standards also 1 mmol/L stock solutions were prepared. The solutions were prepared on melting ice, protected from light, aliquoted in amber glass vials with a Teflon screw cap, and stored at −80 °C. The work solutions of standards and internal standards were freshly prepared in water per run. For internal standards a mix was used for spiking in the samples and calibration curves: 0.1 μmol/L for 5-HT-d4, 5-HTP-d4, 5-HIAA-13C6, KYN-d4, and 1 μmol/L for TRP-d5.
The work solutions for 5-HT, 5-HTP, 5-HIAA, and KYN were 100 μmol/L and 1 μmol/L, WS1 and WS2, respectively. For TRP these were 1 mmol/L and 20 μmol/L, respectively. From these work solutions a combined dilution-series was prepared in water using amber tubes with a final volume of 1 mL. The dilution-series contained 0–0.004–0.01–0.02–0.1–0.2–1–4–6–8 and 10 μmol/L for 5-HT, 5-HTP, 5-HIAA, and KYN. For TRP this was 0–0.1–0.3–0.6–3–6–30–60–100–120 and 150 μmol/L.
Serum samples
Pooled serum samples for validation
Residual serum specimens from daily routine analyses in our laboratory, stored at 5 °C for a maximum of two days, were anonymized and used for preparation of calibration curves in serum matrix and quality control (QC) samples. The serum samples were pooled in culture tubes (50 mL per tube), immediately aliquoted into amber Micronic tubes. For the freeze-thaw cycle testing 1.5 mL of the QC-samples was pipetted in 2 mL Eppendorf tubes. The serum samples were stored at −80 °C.
Serum samples in the first trimester of pregnancy
Serum samples were used from pregnant women (median age (IQR) = 32 ± 6 years) participating in the Rotterdam Periconceptional Cohort (Predict study), a longitudinal prospective periconception hospital-based cohort conducted at the Department of Obstetrics and Gynecology of the Erasmus MC, University Medical Center, Rotterdam, the Netherlands (Erasmus MC). A detailed description of the cohort has been published previously [16, 17]. The Predict study was conducted in accordance with the principles laid down in the Declaration of Helsinki and ethical approval was obtained from the local Medical Ethical Committee of the Erasmus MC, and the Central Committee on Research The Hague, the Netherlands (MEC-2004-227; 15 October 2004) [17]. Informed consent was obtained from all individuals included in this study. Specifically, for application of the validated LC-MS/MS method, serum samples (n=374) were used from a subset of the Predict study including women with placental measurements and anthropometric measurements (n=374). Blood was dawn in the first trimester at the moment of inclusion (median (IQR) = 8 ±2 weeks of pregnancy). After a maximum of four hours in which the blood samples were exposed to room temperature and white light the blood samples were centrifuged at 2,000 g and stored at −80 °C.
Sample preparation
Serum samples were defrosted on melting ice and protected from light. First, 25 µL serum was pipetted in an amber tube. Next, 25 µL internal standard mixture was added and stirred. The following step in sample preparation was protein precipitation using acetonitrile (500 µL). The sample was stirred for 15 s, thereafter kept on melting ice for 10 min, and thoroughly mixed for 5 min using a IKA shaking device. The sample was then centrifuged for 10 min in a cooled (6 °C) benchtop centrifuge at 21,259 g. The supernatant was transferred into an amber Micronic tube and dried at 40 °C under the gentle stream of nitrogen gas. Once the supernatant dried, the sample was dissolved in 100 µL water. The sample was shortly stirred. After being in the IKA shaker for 5 min, the sample was centrifuged using a cooled centrifuge for 2 min at low speed 935 g. Finally, the Micronic tube was transferred to the auto-sampler and 30 µL was injected into the LC-MS system.
Liquid chromatography mass spectrometry
A Waters Xevo-TQ-S micro mass spectrometry system with an ACQUITY I-Class UPLC was used for all experiments (Etten-Leur, the Netherlands). As analytical column a Waters ACQUITY UPLC HSS T3 (2.1 × 100 mm, particle size: 1.8 µm) was used (Etten-Leur, the Netherlands). The analytes were separated by a gradient elution method including water with 0.2% formic acid (eluent A), and acetonitrile with 0.2% formic acid (eluent B) with a flow 0.3 mL/min. The duration of the run was 7 min and all analytes eluted within 5 min. For weak and strong needle washes water and methanol were used, respectively. The auto-sampler and column temperature were 6 and 30 °C. For the LC-MS/MS measurement an optimized multiple reaction monitoring mode (MRM) program in ESI+ was used. For optimal MRM transition the standards and internal standards were tuned by direct infusion and/or combined method, IntelliStart method and LC-MS tuning. The transitions were chosen taking into account the following parameters: maximum intensity of the signal, lowest insource fragmentation, and minimal interferences. The MassLynx version 4.2 was used as a software package, and the data processing was performed by TargetLynx XS (Waters). Table 1 and Supplementary Tables 1 and 2 show the optimal conditions for measurement of the analytes.
Multiple reaction monitoring mode parameters [M+H]+.
| Compound | Parent | Daughter | Cone, volt | Collision, volt |
|---|---|---|---|---|
| Tryptophan | 205.1 | 118.1 | 30 | 25 |
| Tryptophan | 205.1a | 146.1 | 30 | 15 |
| Tryptophan-deuterium 5 | 210.1 | 122.1 | 30 | 25 |
| Tryptophan-deuterium 5 | 210.1b | 150.1 | 30 | 15 |
| Kynurenine | 208.9a | 94.2 | 18 | 12 |
| Kynurenine | 208.9 | 146.2 | 18 | 18 |
| Kynurenine-deuterium 4 | 213.2b | 98.2 | 18 | 12 |
| Kynurenine-deuterium 4 | 213.2 | 150.2 | 18 | 18 |
| 5-Hydroxytryptophan | 221.1 | 134.1 | 22 | 23 |
| 5-Hydroxytryptophan | 221.1a | 162.1 | 22 | 15 |
| 5-Hydroxytryptophan-deuterium 4 | 225.1 | 137.1 | 22 | 23 |
| 5-Hydroxytryptophan-deuterium 4 | 225.1b | 165.1 | 22 | 15 |
| 5-Hydroxytryptamine | 177.1a | 115.1 | 23 | 24 |
| 5-Hydroxytryptamine | 177.1 | 132.1 | 23 | 18 |
| 5-Hydroxytryptamine-deuterium 4 | 181.1b | 118.1 | 23 | 24 |
| 5-Hydroxytryptamine-deuterium 4 | 181.1 | 136.1 | 23 | 18 |
| 5-Hydroxyindole acetic acid | 192.1 | 91.1 | 22 | 35 |
| 5-Hydroxyindole acetic acid | 192.1a | 146.1 | 22 | 12 |
| 13C6-5-Hydroxyindole acetic acid | 198.1 | 96.1 | 22 | 35 |
| 13C6-5-Hydroxyindole acetic acid | 198.1b | 152.1 | 22 | 12 |
-
Method event between 1.81 and 4.59 min to the liquid chromatography (to the detector), aquantifier transition analyte, bquantifier transition internal standard.
Validation
During all validation steps, the samples and standard solutions were kept on melting ice and protected from light.
Calibration curves
The concentrations of the standard calibration curve of TRP were 0–0.1–0.3–0.6–3–6–30–60–100–120–150 μmol/L, while those of KYN, 5-HTP, 5-HT, and 5-HIAA were 0–4–10–20–100–200–1,000–4,000–6,000–8,000 and 10,000 nmol/L (for KYN: 0–0.004–0.01–0.02–0.1–0.2–1–4–6–8 and 10 μmol/L) based on previous literature [18], [19], [20], [21], [22], [23], [24]. Since the concentrations of TRP and most analytes in serum were considerably high, calibration curves were prepared in water for calculation of analytes in the samples. For the data processing the TargetLynx XS program was used. Fit Weighting option was set on 1/X.
To evaluate linearity, we made plots including the observed vs. the expected concentrations, and compared calibration functions (regression analysis) including linear and nonlinear terms. The coefficients of determination (R2) were calculated to assess the goodness-of-fit of the calibration function. The same calibration curves were prepared twice in serum matrix to also determine the goodness-of-fit of the calibration function and linearity in the serum matrix.
Lower limit of detection and lower limit of quantitation
Pooled serum used for validation contained the following endogenous concentrations of analytes: TRP (52.6 ± 6.7 μmol/L), KYN (2.4 ± 0.1 μmol/L), 5-HTP (8.7 ± 1.4 nmol/L), 5-HT (507 ± 79 nmol/L) and 5-HIAA (67 ± 1.1 nmol/L). The lower limit of detection (LLOD) and lower limit of quantitation (LLOQ) were the points at which the response exceeded the baseline signal-to-noise by a factor of 3 and 10, respectively. The LLOD was only determined for 5-HTP and 5-HIAA, since the concentrations of the other analytes was so high in serum that there would be almost no serum matrix left after dilution. To accurately determine the LLOD for 5-HTP and 5-HIAA and the LLOQ for all analytes in serum the following five approaches were used: (1) adding an increasing concentration of analytes in low QC-samples (for 5-HTP and 5-HIAA); (2) spiking the same concentration range of analytes in low QC-samples as well as in water (for 5-HTP and 5-HIAA); (3) adding an increasing concentration of isotope labelled internal standard in low QC-samples; (4) extracting different volumes of low QC-samples; and (5) using different volumes of extract.
Accuracy and precision
Three concentration levels of QC-samples were used to cover the calibration range: low QC, medium QC, and high QC. Supplementary Table 3 provides an overview of the analyte concentrations in the QC-samples. The QC-samples were measured five times with internal standards to determine the accuracy and precision, and once without internal standards to evaluate the internal standards with regard to potential inferring peaks. Accuracy was determined by the percentage of the nominal concentration (recovery (%RC)), and the within-run and between-run precision by the coefficient variation (%CV). The total imprecision was calculated by combining the between-run %CV and within-run %CV
Matrix effect
The matrix effect, extraction recovery and overall process efficiency were evaluated by five-point calibration curves, comparing the analyte responses in the standard solution in water (neat standard), in serum matrix spiked before extraction (pre-extraction spike), and in serum matrix spiked after extraction (post-extraction spike). Analyte matrix effects and recoveries were calculated from the peak areas according to an adaptation of the method of Matuszewski, Constanzer [26]:
Matrix effect (%) =
Extraction recovery (%) =
Overall process efficiency (%) =
The matrix effect was then evaluated using different volumes of low QC-samples (5–10–15–20–25 μL) and different volumes of extraction. Low QC serum was extracted in duplicate and the extracts were combined. From this pooled extract different volumes were dried, dissolved in water and measured by LC-MS/MS according to the standard protocol. The following parameters were compared to the standard extraction method: signal-to-noise of qualifier, ion ratio, signal of internal standard, and peak shape.
Stability
The stability of the five analytes in serum was evaluated using low, medium and high QC-samples. The concentrations of analytes in serum samples exposed to standard conditions were compared to those exposed to (1) room temperature and white light for five hours (equals the extraction time in one run); (2) auto-sampler temperature (6 °C) for 24 h; and (3) freeze and thaw (five days). The three stability tests were all performed in duplicate. The analytes in serum were considered stable when %CV was less than 15%, and for LLOQ less than 20%.
Carryover
Blank samples were injected after the highest concentration of calibration curve and each QC-sample to test carryover.
Trueness
The results of the LC-MS/MS method were compared with the results of a LC-MS/MS assay from the Department of Laboratory Medicine, University Medical Center Groningen. Since the volumes of the serum samples were limited, we focused on comparing the following five samples per analyte: low QC-sample, medium QC-sample, high QC-sample and two calibrators (KYN, 5-HTP, 5-HT, and 5-HIAA: 4 and 6 μmol/L; TRP: 60 and 100 μmol/L). The QC-samples and calibrators were included in each run of our study and the concentration ranges of TRP metabolites was smaller in the samples of the Predict study than in the QC-samples and calibrators. For all analytes, the absolute difference in concentrations measured by the two methods was determined, and the relative difference (%) was calculated by dividing the absolute differences by the analyte concentrations measured by our LC-MS/MS method, multiplied by 100%.
Measurement of first trimester serum samples
The validated LC-MS/MS method was applied to serum samples of pregnant women participating in the Predict study (n=374). The serum samples were analyzed in separate runs, 80 samples in each run. At the beginning of the run solvent (water) and standard solution containing a known concentration of analytes and their internal standards (System Suitability Test) were injected in duplicate. A double blank extraction was always performed in duplicate (50 µL water and 500 µL acetonitrile) and analyzed in each run. At the beginning and end of each run calibration curves were injected (in duplicate). To keep track of the accuracy and precision low, medium and high QC-samples in four-fold were extracted and injected during the run, placed randomly in the run.
Statistical analysis
All statistical analyses were carried out using R version 3.5 (R for Windows). The LC-MS/MS method was compared to a LC-MS/MS method of the University Medical Center Groningen using a Passing Bablok regression. The proportions of values below the LLOD and the LLOQ were presented in pie charts. Analyte concentrations were presented as median with interquartile range (IQR). In addition the analyte concentrations were stratified for BMI, and Pearson correlations between the five analytes and BMI were conducted. Since the distribution of 5-HIAA was skewed a (natural) log transformation was performed to obtain a normal distribution. A p-value ≤0.05 was considered statistically significant.
Results
Selectivity
MS conditions were optimized for all analytes. The measured transitions of the analytes including the used cone voltage and collision energy are presented in Table 1. As illustrated by the MRM chromatograms in Figure 2, this LC-MS/MS method provided satisfactory separation of the analytes within a 7 min run.
Chromatogram of quantifier peaks of the five analytes. aPeak from kynurenine.
Validation
Calibration curves
Plots of the measured vs. true analyte concentrations showed a linear relationship (data on request), and the R2 of all five analytes was >0.995. Moreover, the fit of the spline regression was not better than that of the regression with no nonlinear terms, except for KYN (R2=0.998 vs. 0.999). Since the 5-HTP concentration was considerably lower in the samples compared to other four analytes, additional experiments were performed. The mixed analytes curve was generated in serum matrix with the concentration range of 0–2–3–4–10–15–20–50–100 nmol/L compared to the same concentration range of neat standards solutions. The R2 of both curves was >0.995, and the recovery of serum matrix compared to neat standard solutions was 103%. The medium QC was diluted with low QC as follows: 2 times, 8 times, and 16 times with R2>0.995 (Supplementary Figure 1).
Lower limit of detection and lower limit of quantitation
Since the concentration of most analytes were considerably high in serum matrix, the focus was to determine the LLOQ value of the analytes. Only for 5-HTP and 5-HIAA the LLOD values were determined. For 5-HTP the LLOD was 2.6 nmol/L and the LLOQ was 4.0 nmol/L. 5-HIAA had a LLOD of 4.0 nmol/L and a LLOQ of 8.0 nmol/L. The LLOQ of 5-HT, KYN and TRP were 22.0 nmol/L, 0.011 μmol/L and 0.043 μmol/L, respectively. For the LLOD and LLOQ, the ion ratios of the analytes were within 20% for TRP, KYN, 5-HTP, and 5-HT and within 25% for 5-HIAA.
Accuracy and precision
The within-run and between-run accuracies and precisions are shown in Table 2. The within-run and between-run accuracies ranged between 72 and 97%, and 79 and 104%, respectively. The within-run and between-run precisions were all within 15%, except for the between-run precisions of the low QC-samples of 5-HTP and 5-HT, which were both 16%. During the many runs performed to measure the analyte concentrations of the serum samples of the Predict study, the coefficient variation was <15%. These results indicate that the method is accurate and precise for determining the analytes in serum.
Accuracies and precisions.
| Quality control-level | Lowa | Medium | High | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Analyte | Mean (standard deviation) | Coefficient variation | Total coefficient variation | Mean (standard deviation) | Recovery coefficient | Coefficient variation | Total coefficient variation | Mean (standard deviation) | Recovery coefficient | Coefficient variation | Total coefficient variation |
| Within-run (n=5) | |||||||||||
|
|
|||||||||||
| Tryptophan | 46.8 (0.3) | 1 | 13 | 75.5 (2.0) | 96 | 3 | 9 | 104.8 (2.2) | 97 | 2 | 7 |
| Kynurenine | 2.2 (0.1) | 4 | 6 | 2.6 (0.1) | 80 | 3 | 6 | 5.1 (0.1) | 72 | 2 | 6 |
| 5-Hydroxytryptophan | 11 (1) | 7 | 17 | 354 (24) | 86 | 7 | 8 | 3,190 (197) | 79 | 6 | 10 |
| 5-Hydroxytryptamine | 427 (26) | 6 | 17 | 754 (76) | 82 | 10 | 16 | 3,725 (222) | 82 | 6 | 14 |
| 5-Hydroxyindole Acetic acid | 67 (8) | 11 | 11 | 445 (7) | 95 | 2 | 7 | 3,660 (95) | 89 | 3 | 8 |
|
|
|||||||||||
| Between-run (n=5) | |||||||||||
|
|
|||||||||||
| Tryptophan | 52.6 (6.7) | 13 | 13 | 83.8 (6.9) | 104 | 8 | 9 | 111.3 (7.9) | 98 | 7 | 7 |
| Kynurenine | 2.4 (0.1) | 5 | 6 | 2.7 (0.1) | 91 | 5 | 6 | 5.5 (0.3) | 79 | 6 | 6 |
| 5-Hydroxytryptophan | 9 (1) | 16 | 17 | 368 (9) | 90 | 3 | 8 | 3,306 (265) | 82 | 8 | 10 |
| 5-Hydroxytryptamine | 507 (79) | 16 | 17 | 889 (107) | 96 | 12 | 16 | 4,398 (549) | 97 | 13 | 14 |
| 5-Hydroxyindole Acetic acid | 67 (1) | 2 | 11 | 464 (31) | 99 | 7 | 7 | 3,905 (276) | 96 | 7 | 8 |
-
aConcentrations of analytes in the low QC-samples are endogenous. Concentrations of tryptophan and kynurenine are in µmol/L. Concentrations of 5-hydroxytryptophan, 5-hydroxytryptamine and 5-hydroxyindole acetic acid are in nmol/L.
Matrix effect
The matrix effect, extraction recovery and overall process efficiency are summarized in Table 3. The mean regression coefficient for the five analytes using different volumes of serum samples was at least 0.992, suggesting excellent linearity despite the different volumes of serum samples. Moreover, the signal-to-noise ratio of the qualifier, ion ratio, signal of internal standard, and peak shape, were all acceptable for the different volumes of extracts of the five analytes (10–20–30–40–50–100–200–500 μL). The %CV in the whole range of volume extracts was below 13% for all analyses.
Matrix effects, extraction recoveries and process efficiencies of the five analytes.
| Percentage | Tryptophan | Kynurenine | 5-Hydroxytryptophan | 5-Hydroxytryptamine | 5-Hydroxyindole acetic acid |
|---|---|---|---|---|---|
| Matrix effect | 90 | 79 | 110 | 91 | 93 |
| Extraction recovery | 106 | 111 | 95 | 119 | 96 |
| Process efficiency | 95 | 88 | 105 | 109 | 89 |
Stability
The concentrations of analytes in the serum samples did not change considerably after exposure to room temperature and white light for five hours or after being 24 h at auto-sampler temperature (6 °C). Furthermore, the concentrations of analytes in serum remained stable after five freeze-thaw cycles on different days. The %CV of the low, medium and high QC-samples of each analyte was ≤15.5 (TRP: ≤10%, KYN: <15%, 5-HTP: ≤15%, 5-HT: <15.5%, 5-HIAA: <10%).
Carryover
As previously described, carryover was determined by injecting blank samples after a high concentration sample (highest point of calibration curves): 150 μmol/L for TRP, and 10 μmol/L for KYN, 5-HTP, 5-HT, and 5-HIAA. The analyte responses were for the low, medium and high QC-samples <0.5%, indicating negligible carryover.
Trueness
The results of the LC-MS/MS method were in agreement with those of the LC-MS/MS method of the University Medical Center Groningen. Supplementary Table 4 shows the absolute and relative differences in measured analyte concentrations between the two methods. The relative difference ranged from 0.1 to 26.3% (Supplementary Table 4).
Measurement of first trimester serum samples
Using the validated LC-MS/MS method analytes were determined in a subcohort of the Predict study comprising serum samples (n=374) obtained in the first trimester of pregnancy (median (IQR) = 8 ± 2 weeks of pregnancy). The pie charts in Figure 3 show that the wide range of TRP metabolite concentrations in this study population could be accurately quantified. Table 4 displays the analyte concentrations in the study population, also stratified for BMI. Most women were overweight or obese (53%), followed by women with a BMI within the normal range (47%), and underweight women (1%). The results of the correlation analysis are summarized in Figure 4. A negative correlation was found between BMI and TRP, 5-HTP, and the natural log of 5-HIAA (TRP: r=−0.18, p<0.001; 5-HTP: r=−0.13, p=0.02; natural log of 5-HIAA: r=−0.11, p=0.04). Even though not statistically significant, the same trend was found for 5-HT (r=−0.07, p=0.15). In contrast, BMI was positively associated with KYN (r=0.11, p=0.04).
Proportion of serum samples of the Predict study with analyte concentrations above and below the lower limit of detection and lower limit of quantitation.
Analyte concentrations in the total study population and stratified for body mass index.
| Concentrations | |||||
|---|---|---|---|---|---|
| Body mass index, kg/m2 | Total study population (n=374) | <18 (n=4) | 18–25 (n=174) | 25–30 (n=119) | ≥30 (n=77) |
| Tryptophan | 57.5 (33.1, 101.0) | 71.1 (13.4) | 59.0 (11.6) | 55.8 (15.0) | 55.1 (12.5) |
| Kynurenine | 1.4 (0.7, 3.0) | 1.7 (0.2) | 1.4 (0.3) | 1.4 (0.3) | 1.5 (0.5) |
| 5-Hydroxytryptophan | 4.1 (1.7, 23.1) | 5.4 (1.4) | 4.3 (1.2) | 4.1 (1.1) | 4.2 (1.2) |
| 5-Hydroxytryptamine | 615.0 (16.1, 1960.0) | 552 (87.7) | 639 (343.5) | 614 (296.5) | 563 (347.3) |
| 5-Hydroxyindole acetic acid | 39.9 (18.3, 271.0) | 39.4 (35.5) | 42.3 (20.0) | 38.1 (13.1) | 38.1 (13.6) |
-
The analytes are presented as median with interquartile range. For the total study population the minimum and maximum concentrations are presented. Concentrations of tryptophan and kynurenine are in µmol/L. Concentrations of 5-hydroxytryptophan, 5-hydroxytryptamine and 5-hydroxyindole acetic acid are in nmol/L.
Pearson correlations between body mass index and the five analytes. Correlation analyses were performed with and without outlying values, with outliers defined as values deviating >3 standard deviations from the median. Since, the results of both analyses were comparable only the analysis without outliers were displayed.
Discussion
In this study, we developed and validated an accurate and robust LC-MS/MS method for the simultaneous quantitative analysis of five important TRP metabolites within the kynurenine- and serotonin pathway. Even though the number of LC-MS/MS methods for the analysis of TRP metabolites validated in human serum was limited, the serum concentrations of most TRP metabolites measured in our study were comparable to those reported previously [10], [11], [12], [13], [14]. However, very few studies reported 5-HTP and 5-HIAA concentrations, and the results were inconsistent [21, 23, 24]. Our LC-MS/MS method is unique since it determines TRP metabolites of both the serotonin pathway (5-HTP, 5-HT, and 5-HIAA) and the kynurenine pathway (TRP and KYN). Furthermore, our LC-MS/MS method was very sensitive and only required 25 µL serum to quantify TRP metabolites, while other LC-MS/MS methods required 95–100 µL serum [10, 12, 13]. Finally, the chromatographic separation of our LC-MS/MS method was very efficient as the total run-time was 7 min. This run-time was comparable to that of Tömösi et al. [10] and Hu et al. [13], but twice as short than other LC-MS/MS methods [12, 14].
This validated LC-MS/MS method was applied to serum samples of women in the first trimester of pregnancy, and showed a negative correlation with TRP and serotonin pathway metabolites, and a positive correlation with KYN. An important finding was that BMI correlated negatively with tryptophan and serotonin pathway metabolites and positively with KYN. A possible explanation for the observed correlations might be increased activation of the kynurenine pathway, which is regulated primarily by IDO during pregnancy [27]. An increased BMI is accompanied by low-grade inflammation [28]. The release of cytokines in this pro-inflammatory state can increase the expression and activity of IDO resulting in increased activation of the kynurenine pathway [29, 30]. This may lead to diversion of TRP into the kynurenine pathway and away from the serotonin pathway. This hypothesis is supported by the study of Groer, Fuchs [30] which found a positive association between obesity and kynurenine pathway metabolites in the second and third trimester of pregnancy.
Our validated LC-MS/MS method can be used in future research on the role of TRP metabolites in the (patho)physiology of pregnancy, since it is sensitive, able to quantitate five TRP metabolites from the serotonin and kynurenine pathways in a single short run, and requires only a small amount of serum. Especially studies performed in women of reproductive age and during pregnancy remain of interest with regard to prevention, since many of the pregnancy complications originate in the periconception period, covering the preconception period and early pregnancy [31].
Acknowledgments
We wish to thank all women of the Rotterdam Periconceptional cohort (Predict study) for their participation, and the researchers of the Predict study for recruiting participants and collecting data. In addition, we gratefully acknowledge the help provided by Wim Schilleman, Medical Analyst of the Department Clinical Chemistry, Erasmus MC, University Medical Center, Rotterdam, the Netherlands.
-
Research funding: None declared.
-
Author contributions: All authors have accepted responsibility for the entire content of this manuscript and approved its submission.
-
Competing interests: Authors state no conflict of interest.
-
Informed consent: Informed consent was obtained from all individuals included in this study.
-
Ethical approval: Research involving human subjects complied with all relevant national regulations, institutional policies and is in accordance with the tenets of the Helsinki Declaration (as revised in 2013), and has been approved by the local Medical Ethical Committee of the Erasmus MC, and the Central Committee on Research The Hague, the Netherlands (MEC-2004-227; 15 October 2004).
References
1. Alkhalaf, LM, Ryan, KS. Biosynthetic manipulation of tryptophan in bacteria: pathways and mechanisms. Chem Biol 2015;22:317–28. https://doi.org/10.1016/j.chembiol.2015.02.005.Search in Google Scholar PubMed
2. Roth, W, Zadeh, K, Vekariya, R, Ge, Y, Mohamadzadeh, M. Tryptophan metabolism and gut-brain homeostasis. Int J Mol Sci 2021;22:1–23. https://doi.org/10.3390/ijms22062973.Search in Google Scholar PubMed PubMed Central
3. Broekhuizen, M, Klein, T, Hitzerd, E, de Rijke, YB, Schoenmakers, S, Sedlmayr, P, et al.. l-tryptophan-induced vasodilation is enhanced in preeclampsia: studies on its uptake and metabolism in the human placenta. Hypertension 2020;76:184–94. https://doi.org/10.1161/hypertensionaha.120.14970.Search in Google Scholar
4. Badawy, AA. Kynurenine pathway of tryptophan metabolism: regulatory and functional aspects. Int J Tryptophan Res 2017;10:1–20. https://doi.org/10.1177/1178646917691938.Search in Google Scholar PubMed PubMed Central
5. Tóth, F, Cseh, EK, Vécsei, L. Natural molecules and neuroprotection: kynurenic acid, pantethine and α-lipoic acid. Int J Mol Sci 2021;22:1–25. https://doi.org/10.3390/ijms22010403.Search in Google Scholar PubMed PubMed Central
6. Gao, K, Mu, CL, Farzi, A, Zhu, WY. Tryptophan metabolism: a link between the gut microbiota and brain. Adv Nutr 2020;11:709–23. https://doi.org/10.1093/advances/nmz127.Search in Google Scholar PubMed PubMed Central
7. Gumusoglu, S, Scroggins, S, Vignato, J, Santillan, D, Santillan, M. The serotonin-immune axis in preeclampsia. Curr Hypertens Rep 2021;23:37. https://doi.org/10.1007/s11906-021-01155-4.Search in Google Scholar PubMed PubMed Central
8. Keaton, SA, Heilman, P, Bryleva, EY, Madaj, Z, Krzyzanowski, S, Grit, J, et al.. Altered tryptophan catabolism in placentas from women with pre-eclampsia. Int J Tryptophan Res 2019;12:1–8. https://doi.org/10.1177/1178646919840321.Search in Google Scholar PubMed PubMed Central
9. Whiley, L, Nye, LC, Grant, I, Andreas, N, Chappell, KE, Sarafian, MH, et al.. Ultrahigh-performance liquid chromatography tandem mass spectrometry with electrospray ionization quantification of tryptophan metabolites and markers of gut health in serum and plasma-application to clinical and epidemiology cohorts. Anal Chem 2019;91:5207–16. https://doi.org/10.1021/acs.analchem.8b05884.Search in Google Scholar PubMed PubMed Central
10. Tömösi, F, Kecskeméti, G, Cseh, EK, Szabó, E, Rajda, C, Kormány, R, et al.. A validated UHPLC-MS method for tryptophan metabolites: application in the diagnosis of multiple sclerosis. J Pharm Biomed Anal 2020;185:1–12. https://doi.org/10.1016/j.jpba.2020.113246.Search in Google Scholar PubMed
11. Zhu, W, Stevens, AP, Dettmer, K, Gottfried, E, Hoves, S, Kreutz, M, et al.. Quantitative profiling of tryptophan metabolites in serum, urine, and cell culture supernatants by liquid chromatography-tandem mass spectrometry. Anal Bioanal Chem 2011;401:3249–61. https://doi.org/10.1007/s00216-011-5436-y.Search in Google Scholar PubMed
12. Sadok, I, Jędruchniewicz, K, Rawicz-Pruszyński, K, Staniszewska, M. UHPLC-ESI-MS/MS quantification of relevant substrates and metabolites of the kynurenine pathway present in serum and peritoneal fluid from gastric cancer patients-method development and validation. Int J Mol Sci 2021;22:1–21. https://doi.org/10.3390/ijms22136972.Search in Google Scholar PubMed PubMed Central
13. Hu, LJ, Li, XF, Hu, JQ, Ni, XJ, Lu, HY, Wang, JJ, et al.. A simple HPLC-MS/MS method for determination of tryptophan, kynurenine and kynurenic acid in human serum and its potential for monitoring antidepressant therapy. J Anal Toxicol 2017;41:37–44. https://doi.org/10.1093/jat/bkw071.Search in Google Scholar PubMed
14. Takahashi, S, Iizuka, H, Kuwabara, R, Naito, Y, Sakamoto, T, Miyagi, A, et al.. Determination of l-tryptophan and l-kynurenine derivatized with (R)-4-(3-isothiocyanatopyrrolidin-1-yl)-7-(N, N-dimethylaminosulfonyl)-2, 1, 3-benzoxadiazole by LC-MS/MS on a triazole-bonded column and their quantification in human serum. Biomed Chromatogr 2016;30:1481–6. https://doi.org/10.1002/bmc.3709.Search in Google Scholar PubMed
15. Favennec, M, Hennart, B, Caiazzo, R, Leloire, A, Yengo, L, Verbanck, M, et al.. The kynurenine pathway is activated in human obesity and shifted toward kynurenine monooxygenase activation. Obesity 2015;23:2066–74. https://doi.org/10.1002/oby.21199.Search in Google Scholar PubMed
16. Rousian, M, Schoenmakers, S, Eggink, AJ, Gootjes, DV, Koning, AH, Koster, MP, et al.. Cohort profile update: the Rotterdam periconceptional cohort and embryonic and fetal measurements using 3D ultrasound and virtual reality techniques. Int J Epidemiol 2021;50:1426–7. https://doi.org/10.1093/ije/dyab030.Search in Google Scholar PubMed PubMed Central
17. Steegers-Theunissen, RP, Verheijden-Paulissen, JJ, van Uitert, EM, Wildhagen, MF, Exalto, N, Koning, AH, et al.. Cohort profile: the Rotterdam periconceptional cohort (Predict study). Int J Epidemiol 2016;45:374–81. https://doi.org/10.1093/ije/dyv147.Search in Google Scholar PubMed
18. Sha, Q, Madaj, Z, Keaton, S, Escobar Galvis, ML, Smart, L, Krzyzanowski, S, et al.. Cytokines and tryptophan metabolites can predict depressive symptoms in pregnancy. Transl Psychiatry 2022;12:1–8. https://doi.org/10.1038/s41398-022-01801-8.Search in Google Scholar PubMed PubMed Central
19. Nilsen, RM, Bjørke-Monsen, AL, Midttun, O, Nygård, O, Pedersen, ER, Ulvik, A, et al.. Maternal tryptophan and kynurenine pathway metabolites and risk of preeclampsia. Obstet Gynecol 2012;119:1243–50. https://doi.org/10.1097/aog.0b013e318255004e.Search in Google Scholar PubMed PubMed Central
20. Cengiz, H, Dagdeviren, H, Caypinar, SS, Kanawati, A, Yildiz, S, Ekin, M. Plasma serotonin levels are elevated in pregnant women with hyperemesis gravidarum. Arch Gynecol Obstet 2015;291:1271–6. https://doi.org/10.1007/s00404-014-3572-2.Search in Google Scholar PubMed
21. Boulet, L, Faure, P, Flore, P, Montérémal, J, Ducros, V. Simultaneous determination of tryptophan and 8 metabolites in human plasma by liquid chromatography/tandem mass spectrometry. J Chromatogr B: Anal Technol Biomed Life Sci 2017;1054:36–43. https://doi.org/10.1016/j.jchromb.2017.04.010.Search in Google Scholar PubMed
22. Comai, S, Bertazzo, A, Carretti, N, Podfigurna-Stopa, A, Luisi, S, Costa, CV. Serum levels of tryptophan, 5-hydroxytryptophan and serotonin in patients affected with different forms of amenorrhea. Int J Tryptophan Res 2010;3:69–75. https://doi.org/10.4137/ijtr.s3804.Search in Google Scholar PubMed PubMed Central
23. van Faassen, M, Bouma, G, de Hosson, LD, Peters, MA, Kats-Ugurlu, G, de Vries, EG, et al.. Quantitative profiling of platelet-rich plasma indole markers by direct-matrix derivatization combined with LC-MS/MS in patients with neuroendocrine tumors. Clin Chem 2019;65:1388–96. https://doi.org/10.1373/clinchem.2019.305359.Search in Google Scholar PubMed
24. Carretti, N, Bertazzo, A, Comai, S, Costa, CV, Allegri, G, Petraglia, F. Serum tryptophan and 5-hydroxytryptophan at birth and during post-partum days. Adv Exp Med Biol 2003;527:757–60.10.1007/978-1-4615-0135-0_90Search in Google Scholar PubMed
25. Chace, DH, Barr, JR, Duncan, MW, Matern, D, Morris, MR, Palmer-Toy, DE, et al.. Mass spectrometry in the clinical laboratory: general principles and guidance; approved guideline. Clinical and Laboratory Standards Institute; 2007. 1–97 pp.Search in Google Scholar
26. Matuszewski, BK, Constanzer, ML, Chavez-Eng, CM. Strategies for the assessment of matrix effect in quantitative bioanalytical methods based on HPLC-MS/MS. Anal Chem 2003;75:3019–30. https://doi.org/10.1021/ac020361s.Search in Google Scholar PubMed
27. Badawy, AA. Tryptophan metabolism, disposition and utilization in pregnancy. Biosci Rep 2015;35:1–16. https://doi.org/10.1042/bsr20150197.Search in Google Scholar
28. Saltiel, AR, Olefsky, JM. Inflammatory mechanisms linking obesity and metabolic disease. J Clin Invest 2017;127:1–4. https://doi.org/10.1172/jci92035.Search in Google Scholar
29. Badawy, AA. Plasma free tryptophan revisited: what you need to know and do before measuring it. J Psychopharmacol 2010;24:809–15. https://doi.org/10.1177/0269881108098965.Search in Google Scholar PubMed
30. Groer, M, Fuchs, D, Duffy, A, Louis-Jacques, A, D’Agata, A, Postolache, TT. Associations among obesity, inflammation, and tryptophan catabolism in pregnancy. Biol Res Nurs 2018;20:284–91. https://doi.org/10.1177/1099800417738363.Search in Google Scholar PubMed PubMed Central
31. Steegers-Theunissen, RP, Twigt, J, Pestinger, V, Sinclair, KD. The periconceptional period, reproduction and long-term health of offspring: the importance of one-carbon metabolism. Hum Reprod Update 2013;19:640–55. https://doi.org/10.1093/humupd/dmt041.Search in Google Scholar PubMed
Supplementary Material
The online version of this article offers supplementary material (https://doi.org/10.1515/cclm-2022-0790).
© 2022 the author(s), published by De Gruyter, Berlin/Boston
This work is licensed under the Creative Commons Attribution 4.0 International License.
