Abstract
Tryptophan (TRP) is an essential amino acid catabolized mainly through the kynurenine pathway, and part of it is catabolized in the brain. The abnormal depletion of TRP and production of kynurenine (KYN) by two enzymes, tryptophan 2,3-dioxygenase (TDO) and indoleamine 2,3-dioxygenase (IDO), have been linked to various neurological diseases. The ratio of TRP/KYN in plasma is a valuable measure for IDO/TDO activity and the prognosis of disease conditions. The 4-vinylphenylboronic acid (4-VPBA) was evaluated as a novel stationary phase for OT–CEC–MS/MS. TRP, KYN, and 3-hydroxykynurenine were separated using optimum conditions of 15 mM (NH4)2CO3 at pH 8 as a background electrolyte and 25 kV separation voltage on a 90 cm column. The usefulness of the 4-VPBA column for simple, fast, repeatable, and sensitive CEC–ESI–MS/MS application was demonstrated for the quantitation of TRP and KYN in the plasma of healthy human subjects and neuroinflammation subjects. The plasma sample was extracted on a zirconia-based ion-exchange cartridge for simultaneous protein precipitation and phospholipid removal. The method of standard addition, in combination with the internal standards approach, was used to prepare the calibration curve to overcome matrix matching and eliminate procedural errors. The developed quantitation method was validated according to FDA guidelines for sensitivity, accuracy, precision, and extraction recovery. The measured plasma level of TRP and KYN in healthy humans is aligned with the human metabolome database for the same two metabolites.
Keywords: kynurenines, neuroinflammation, open tubular capillary electrochromatography–mass spectrometry, solid–liquid extraction, tryptophan, The Human Metabolome Database
1 |. INTRODUCTION
Tryptophan (TRP) is one of the essential amino acids obtained from dietary sources [1, 2]. The majority of TRP (~80%) is metabolized via the kynurenine pathway (KP), whereas the remaining portion is utilized in the central nervous system for the synthesis of the neurotransmitter serotonin and melatonin [2, 3]. The KP is associated with many disease conditions. For example, Alzheimer’s disease, human immune deficiency virus, Huntington’s disease, amyotrophic lateral sclerosis, depression, schizophrenia, and cancer [4–6]. The degradation of TRP via KP is initiated by two rate-limiting enzymes, tryptophan 2,3-dioxygenase (TDO) in the liver and indoleamine 2,3-dioxygenase (IDO) in the central nervous system, placenta, and lungs [7]. These two enzymes, TDO and IDO, are known to be activated or overactivated by inflammatory molecules such as cytokines (interferons, e.g., tumor necrosis factor-alpha [TNF-α], interleukins) [8–10]. The catabolism of TRP by two enzymes produces kynurenine (KYN), which produces metabolic products, some neurotoxic and some neuroprotective [3]. Simultaneous measurement of TRP and KYN (i.e., KYN/TRP) ratio can be used to determine TDO and IDO activities in disease conditions.
Various analytical techniques have been developed for measuring TRP and its metabolites in biological matrices for several decades. Examples of such methods are high-performance liquid chromatography (HPLC), GC, and CE. More specifically, mass spectrometric (MS) detection in hyphenation with the separation techniques mentioned before has become the most sensitive and valuable. Numerous studies are recently reported in the literature, and the majority of them reported the use of HPLC–MS/MS [4, 11–14], but there are few reports on GC–MS/MS [15] and CE–MS/MS [16, 17] for separation and quantitation of TRP and its metabolites.
Capillary electrochromatography (CEC) combined with MS/MS is a newly emerging metabolic platform with great potential for separating and analyzing complex metabolites in biological samples [18], including specimens, such as brain tissues, cancer cells, and so forth [19, 20]. The eases of designing exotic stationary phases for separating polar, nonpolar, ionic, and chiral compounds are some of the common advantages of CEC, but the nanoliter (nL) injection volume, very low column pressure, and high efficiency are some unique features of CEC–MS/MS over capillary HPLC–MS/MS. Although reducing internal column diameter using nano-HPLC–MS allows nL injection and improves analytical figures of merit, the latter benefit is only applied for limited cases.
The successful and sensitive quantitation of biological samples by HPLC–MS or CE–MS requires efficient cleanup of interfering proteins and lipids. The presence of proteins and lipids can cause ion suppression during ESI, reduce sensitivity, and contaminate the mass spectrometer’s ion source [21]. In addition, the reproducibilities of electroosmotic flow (EOF) and retention time stability in CEC–MS are significantly affected by the adsorption of proteins and lipids. The conventional protein precipitation method is the most commonly used cleanup method for plasma samples. However, removing phospholipids is often timeconsuming and inefficient, contributing to ion suppression [22]. Hence, there is a need to develop a reliable and faster extraction method for removing protein and phospholipids for improved quantitation in challenging biological matrices. Reports on internal standardization, standard addition, and external matrix-matched calibration are abundant in the literature [23]. However, matrix matching is sometimes tricky as an interference-free matrix is not available in all cases.
In this study, we report using 4-vinylphenylboronic acid (4-VPBA) as a novel stationary phase for CEC–MS/MS application to separate TRP, KYN, and 3-hydroxykynurenine (3-OHKYN). Several chromatographic selectivity parameters were optimized, such as the concentration of 4-VPBA in the polymerization mixture, pH of background electrolyte (BGE), separation voltage, and column length. Next, the CEC–MS method was validated for accurate measurement of TRP and KYN in a healthy human plasma sample that is critical for monitoring the activity of IDO/TDO. Furthermore, to improve the quantitation of TRP and KYN in human plasma, we combined liquid–liquid extraction with the solid phase ion-exchange method for plasma sample cleanup before ESI–MS detection. To ensure adequate selectivity and correct for matrix effect, we proposed the combined use of standard addition and internal standard (IS) method for the quantitation of endogenous TRP and KYN in human plasma.
2 |. MATERIALS AND METHODS
2.1 |. Chemicals and reagents
l-Tryptophan (TRP), l-kynurenine (98%) (KYN), 3-hydroxy-dl-kynurenine (98%) (3-OHKYN), kynurenic acid (KA), l-tyrosine (99+%) (TYR), 4-VPBA, anhydrous methanol (99.8%), [ϒ-(methacryloyloxy)propyl]trimethoxysilane (98.0%) (ϒ-MAPS), sodium hydroxide (NaOH, 50% v/v), acetic acid (99.7%), formic acid (>95%), and healthy human plasma (lyophilized powder containing 4% trisodium citrate anticoagulant) were purchased from Sigma Aldrich. 5-Hydroxy-l-tryptophan (5-OHTRP) was purchased from MP Biomedicals LLC, Solon, OH, USA. Stable isotopic labeled IS (SIL-IS) l-tryptophan-D8 was purchased from CDN isotopes. HPLC-grade methanol (99.9%) and acetonitrile (99.9%) were purchased from Fischer Chemicals. Ammonium acetate (28%–30% w/w) and 14.8 M ammonium hydroxide (28%–30%) were purchased from EMD chemicals. Triple deionized (TDI) water was obtained from the Barnstead Nanopure II water system (Barnstead International, Dubuque, IA, USA). Phree Phospholipid removal tubes were purchased from Phenomenex.
2.2 |. OT–CEC–ESI–MS/MS
2.2.1 |. Preparation of CEC column
The two steps for 4-VPBA column preparation are shown in Figure 1. The capillary bonding was carried out by slight modification in the procedure reported elsewhere [18]. First, the fused silica (FS) capillary (75 cm length, 50 μm i.d., 360 μm o.d.) was conditioned by flushing sequentially with TDI water (30 min), 1 M sodium hydroxide (3 h), TDI water (30 min), 0.1 M HCl (30 min), TDI water (30 min), and anhydrous methanol (30 min). Next, the pretreated capillary was flushed for 30 min with (1:1) ϒ-MAPS/MeOH solution (prepared by ultra-sonication for 10 min). Both the capillary ends were sealed with rubber septa and heated at 60°C (20 h). The capillary was flushed with anhydrous methanol (30 min) and dried with nitrogen for 3 h at 60°C. The solution was prepared by dissolving 25, 30, 40, 50, and 70 mg of 4-VPBA in 2.5 mL anhydrous methanol and sonicating for 10 min. The coating solution was then flushed through the capillary for 30 min. Both capillary ends were sealed, and the capillary was heated in a GC oven at 60°C for 24 h. The capillary was flushed with methanol for 30 min to remove unreacted 4-VPBA and dried with nitrogen at 60°C for 3 h.
FIGURE 1.
Schematic for the preparation of 4-vinylphenylboronic acid (4-VPBA) covalently bonded column.
2.2.2 |. Preparation of sheath liquid and background electrolyte (BGE)
Sheath liquid for MS ionization (40 mM HOAc in 80/20 [v/v] MeOH/H2O) was prepared by adding 460 μL of 17.4 M HOAc to 160 mL of MeOH and adjusting the final volume to 200 mL with TDI H2O. The prepared sheath liquid was filtered with a 0.2 μm nylon membrane filter and degassed for 20–25 min before its use. The BGE was prepared by dissolving an appropriate amount of NH4CO3 in 80 mL of TDI water and adjusting the pH with 17.4 M HOAc or 14.8 M NH4OH using an Orion 420A pH meter (Beverly, MA, USA). The final buffer volume was adjusted to 100 mL with TDI water. The BGE was filtered through a 0.45 μm nylon membrane filter and degassed for 10–15 min before use.
2.2.3 |. CE–MS instrumentation
All CEC–ESI–MS/MS experiments were performed on Agilent 7100 Capillary Electrophoresis system coupled with an Agilent 6410B triple quadrupole mass spectrometer equipped with an Agilent CE–MS adapter kit (G-1603A) and an Agilent CE–ESI–MS sprayer kit (G-1607). An 1100 Agilent isocratic LC pump supplied the sheath liquid for ESI. The Agilent MassHunter Workstation Data Acquisition software (B.08.00 version) was used to control the instrument, and the MassHunter Qualitative Analysis software (B.07.00 version) was used for chromatographic data analysis. For quantitative analysis, the peak area of chromatograms was calculated using the Agile2 integrator algorithm. Flow injection analysis for multiple reaction monitoring (MRM) parameters optimizations was carried out on Agilent 1100 series HPLC equipped with a binary pump (including degasser), a thermostatted autosampler, and a column compartment (temperature controlled). The Agilent Optimizer software was used to optimize MRM parameters, such as fragmentor voltage and collision energy for TRP, three metabolites of TRP, and standard isotopic labels (Table S1).
2.2.4 |. Preparation of stock and standard solutions for qualitative analysis
A 5 mM stock solution of TRP, KYN, 3-OHKYN, and 3-OHTRP was prepared in TDI water and stored at 0°C. A new standard solution was prepared, thawing the stock solutions and appropriately aliquoting the volume of stocks daily by diluting them with TDI water. The final mixture was filtered with a 0.2 μm nylon syringe filter.
Preparation of standards, internal standards, and calibrants for quantitation
Stock solutions of TRP and KYN were prepared at 5 mM in 80/20 ACN/H2O, whereas stock solutions of ISs TRP-D8 and TYR were prepared at 5 and 10 mM, respectively, in 80/20 ACN/H2O and stored at 0°C. For method validation, working mixture A1–A5 containing TRP (220, 440, 879, 1759, and 3517 μM) and KYN (13, 25, 50, 101, and 201 μM) were prepared by aliquoting a certain volume from 5 mM TRP and KYN stocks and diluting with 80/20 ACN/H2O. Working IS mixtures B containing TRP-D8 and TYR at 255 μM were prepared by aliquoting 56 and 28 μL of TRP-D8 and TYR, respectively, and diluting to 1100 μL with 80/20 ACN/H2O. Five calibration standard mixtures C1–C5 containing IS at a constant concentration (12 μM TRP-D8 and 12 μM TYR), TRP (209, 419, 837, 1675, and 3350 μM), and KYN (12, 24, 48, 94, and 192 μM) were prepared by adding 20 μL of IS mixture B to 400 μL of each A1–A5.
For the quantitation of the patient plasma samples, the working standard mixture was prepared by serial dilutions at final concentrations of 59, 95, 152, 242.5, and 388 μM TRP; and 4, 6, 9.4, 15, and 24 μM KYN. The working IS mixture D containing 1440 μM TRP-D8 and 480 μM TYR was prepared by aliquoting 288 μL of TRP-D8 and 480 μL of TYR stock solutions and diluting to a final volume of 1 mL. Five calibration standard mixtures E1–E5 containing IS at a constant concentration (90 μM TRP-D8 and 30 μM TYR), TRP (55.5, 89, 142, 227, and 364 μM), and KYN (3.4, 5.5, 9, 14, 22.5 μM) were prepared by adding 40 μL of IS mixture D to each 600 μL of each E1–E5.
2.2.5 |. Plasma sample preparation
Plasma samples of healthy volunteers and patients with low and high inflammation were obtained from Emory University and stored at −80°C. Commercial human plasma (the entire content of the bottle) was dissolved in 5 mL of TDI water as recommended by the supplier and stored at −80°C. All the procedures of plasma sample collections from human subjects at Emory University were approved by Emory Institutional Review Board, and the subjects provided written informed consent.
Calibration curves were prepared by combining standard addition and IS methods using commercial human plasma. Sample cleanup (protein and phospholipid removal) was carried out by solid–liquid extraction (SLE) using Phree phospholipid removal cartridges (Figure S1). Briefly, 300 μL of acetonitrile with 1% formic acid was added to SLE cartridges, followed by the addition of 100 μL of thawed plasma sample and 50 μL of calibration standard mixture (C1–C5 or E1–E5). The SLE cartridges were then vortexed for 5 min at a moderate speed to support mixing and protein precipitation. The extracts were eluted and collected in 1.5 mL Eppendorf tubes by applying a vacuum (10–15 in. Hg) for 5 min. From the collected extracts, 300 μL of eluents from each tube was transferred to new tubes, dried, and reconstituted in 100 μL of TDI water before injection into the CEC column. The unknown healthy plasma sample was prepared by adding 300 μL of acetonitrile containing 1% FA, 100 μL of plasma, and 50 μL of IS mixture B. The unknown patient’s plasma samples were prepared by adding 300 μL of acetonitrile containing 1% FA, 100 μL of plasma, and 50 μL of IS mixture D.
All protocols involving the human subjects from Emory University were a priori approved by the Emory University Institutional Review Board. In addition, all human subjects recruited at Emory University were provided with written informed consent. A total of 34 plasma samples, of which two samples obtained from each subject at an interval of 1–3 h, were collected as blinded samples to the clinical status of study participants.
3 |. RESULTS AND DISCUSSION
The utility of 4-VPBA was evaluated as a novel stationary phase for OT–CEC–MS. First, 4-VPBA concentration in the polymerization solution was tested for the optimum column selectivity of a model test mixture containing TRP and its three metabolites (KYN, 3-OHKYN, 5-OHTRP). Next, separation conditions, such as pH, BGE concentration, separation voltage, and column length, were optimized for the best separation with the shortest run time of three critical KP compounds. The stability of the VPBA column was evaluated for its application in the analysis of plasma samples of depressed patients.
3.1 |. Optimization of VPBA concentration in the polymerization solution
FS capillaries were bonded using 4-VPBA solutions prepared at 25, 30, 40, 50, and 70 mg in 2.5 mL methanol using the procedure outlined in Figure 1. A mixture of four analytes (TRP, OHTRP, KYN, and 3-OHKYN) was tested to determine the monomer concentration needed for polymerization to allow the best resolution with the shortest possible run time using 4-VPBA stationary phase. The optimum stationary phase concentration was determined by comparing resolution and the total run time. Increased retention of all four peaks was observed with an increase in stationary phase concentration (Figure 2, top to bottom chromatograms). At 10 mg/mL 4-VPBA concentration, TRP, KYN, and 5-OHTRP coeluted, and only two out of four peaks were resolved, whereas at 12 mg/mL, the hydroxy derivatives of KYN and TRP are resolved, but critical pair of TRP and KYN are not separated. Further increases in the concentration of 4-VPBA to 16 and 20 mg/mL four peaks were seen, whereas the first two peaks of TRP and KYN remain partially resolved. At the highest stationary phase concentration of 28 mg/mL, some improvement is seen for the TRP/KYN peak pair, the last eluting metabolites (3-OHKYN) showed a significant increase in retention time to ~32 min resulting in poor repeatability >5% RSD for retention time, and the total analysis time of 35 min was substantial. A stationary phase concentration of 16 mg/mL (slightly below optimum) was selected to optimize the separation of TRP and two of its metabolites (KYN and 3-OHKYN) within a reasonable run time, as detailed in the following two subsections.
FIGURE 2.
Effect of vinylphenylboronic acid (VPBA) concentration in column coating solution on the selectivity of tryptophan (TRP) and its three metabolites. VPBA-bonded capillary: 50 μm i.d., 67.5 cm total length. Experimental conditions: background electrolyte (BGE): 15 mM (NH4)2CO3, pH 11.4 (adjusted by 16.5 M NH4OH), injection: 5 mbar for 100 s, applied voltage: +15 kV. Mass spectrometric (MS) spray chamber parameters: sheath liquid: 80/20 MeOH/H2O, 40 mM HOAc, sheath liquid flow rate: 5 μL/min, DGT: 200°C, DGF: 5 L/min, nebulizer pressure: 7 psi, ESI capillary voltage: +3500 V. Analytes concentration: 10 μM TRP, 10 μM kynurenine (KYN), 20 μM 5-hydroxytryptophan (5-OHTRP), 30 μM 3-hydroxykynurenine (3-OHKYN). Peak identification: (a) TRP, (b) KYN, (c) 5-OHTRP, and (d) 3-OHKYN. For multiple reaction monitoring (MRM) parameters see Table S1.
3.2 |. Optimization of mobile phase pH and BGE concentration
Initially, separation was attempted for TRP, 5-OHTRP, KYN, and 3-OHKYN at high BGE pH range 10, 10.5, 11, and 11.4 (data not shown). No separation was observed among TRP, KYN, and 5-OHTRP, and only 3-OHKYN was separated from the remaining three mentioned compounds in the highly basic pH of the BGE. However, a partial separation of TRP/KYN was observed only at pH 11.4. Even though higher pH was favorable for the separation of TRP/KYN, the stability of the stationary phase was questionable and found to be compromising the lifetime of the 4-VPBA column. We suspect that the poor column stability could be due to the slow deterioration of SP at a high pH of the mobile phase (pH 11.3) used for separation. Next, the separation was attempted at a pH range from 4.0 to 8.5 (Figure 3). At pH 4.0, no peak eluted but at pH 5.0, there is substantial peak broadening due to very low EOF on the 4-VPBA column, mainly due to the protonation of boronic acid groups on the 4-VPBA moiety. An increase in EOF at pH 6.0 or greater promotes some deprotonation of 4-VPBA groups promoting elution and sharper bands till a plateau in EOF is reached in the pH range of 7.0–8.0 (for both EOF and chromatographic selectivity). Further, an increase in pH above 8.0 did increase EOF, but chromatographic selectivity dropped, resulting in no resolution at pH 8.0 of the TRP/KYN peak pair, but 3-OHKYN was still baseline separated. The drop in separation selectivity and hence the chromatographic resolution at pH > 8 is unclear at this time. Nevertheless, pH 8.0 was found to be the best possible pH for providing a hint for separating TRP/KYN critical peak pair. Because TRP and OHTRP coeluted at pH 4.0–8.5, the separation selectivity of TRP/OHTRP was not further optimized. The EOF profiles of uncoated and coated columns were measured and compared using thiourea as the EOF marker peak (Figure 3B). The EOF profile of coated column and bare silica was similar; however, the magnitude of EOF was reduced on a coated column. The reduction of EOF demonstrated a successful modification of the capillary wall. The relative retention (tR/t0, Figure 3C) profiles of TRP, KYN, and 3-OHKYN were evaluated in the pH range of 5.0–8.5, which further suggests that the maximum difference in tR/t0 values for all three compounds is best obtained at pH8.0.
FIGURE 3.
(A) Effect of background electrolyte (BGE) pH on chromatographic separation of tryptophan (TRP) and its metabolites. Experimental conditions: BGE: 15 mM (NH4)2CO3, pH 4.0–8.5 (adjusted by 17.4 M HOAc), injection: 5 mbar for 100 s, applied voltage: +15 kV, vinylphenylboronic acid (VPBA) column: 50 μm (i.d.) 67.5 cm (length), mass spectrometric (MS) spray chamber parameters are the same as in Figure 2. Analytes concentration: 40 μM TRP, 60 μM kynurenine (KYN), 120 μM 3-hydroxykynurenine (3-OHKYN). Peak identification: (a) TRP, (b) KYN, and (c) 3-OHKYN. (B) Comparison of electroosmotic flow (EOF) profiles at a pH range of 5.0–8.5 on bare fused silica (BFS) versus 4-VPBA columns. Experimental conditions are the same as (A) except for BGE pH 4.0–8.5 on the BFS column and pH 5.0–8.5 on the 4-VPBA column. (C) Relative retention (tR/t0) as a function of pH. Experimental conditions are the same as (A) except BGE pH 5.0–8.5, sample concentration: 40 μM TRP, 40 μM OHTRP, 80 μM KYN, and 120 μM OHKYN. Thiourea (5 mM dissolved in triple deionized [TDI] H2O) was used as a neutral marker to measure t0 and, consequently, the EOF.
The robustness of the CEC–MS column is crucial for developing a repeatable assay in a biospecimen. The 4-VPBA column was tested at the optimized mobile phase pH of 8.0. As shown in Figure S2, a very stable retention time was obtained for 40 consecutive runs in a single day for all three testing compounds. The %RSD of retention times and the resolutions (TRP/KYN and KYN/3-OHKYN) are shown in the inset table of Figure S2. The precision of retention times (i.e., RSD of 2%) for all three analytes showed an excellent stability of the column. The repeatability of the resolution was found to be acceptable for the first peak pair (TRP/KYN). The average resolution for KYN/3-OHKYN was 2.9, but the %RSD of (KYN/3-OHKYN) resolution was as high as 13%. The intra-column and intercolumn repeatability were further evaluated for columns prepared on day 2. The average retention times and %RSD for 30 consecutive injections for two columns are listed in Table S2. The %RSD of retention time for three analytes in column 1 was 2%, whereas the %RSD for peak areas ranged from 11% to 14%. In column 2, the %RSD of retention time and peak area was between 5%–6% and 6%–15%, respectively. The intercolumn precision of retention time for three analytes was between 4%–5% RSD and 8%–9% for the peak area of TRP and KYN. Perhaps, due to the longer retention time of 3-OHKYN, the intercolumn peak area repeatability was high (41% RSD).
3.3 |. Optimization of column length and field strength
Next, the effect of column length at various electric field (E) strengths of 125,167, 208, and 250 kV/cm on TRP/KYN critical pair resolution was studied. TRP, KYN, and 3-OHKYN were separated at three different column lengths, 67.5, 90, and 120 cm at equivalent E. Several trends were observed when Rs and analysis time were plotted as a function of E (Figure 4A,B). First, note the Rs of the TRP/KYN pair decrease on the shortest (67.5 cm) and the longest (120 cm) CEC columns with increasing E. The increased joule heating on the shortest column decreases efficiency, decreasing Rs. On the other hand, longitudinal diffusion (due to longer run times) reduces the peak width on the longest 120 cm column. Second, the column length of 90 cm showed an increase in Rs with increasing E, with the highest Rs of TRP/KYN obtained at 208 kV/cm. Further rise in E does not affect the Rs of the above pair. The total analysis time (i.e., the retention of the last peak) was plotted against V/cm for three column lengths (Figure 4B), and as expected, analysis time was longer on the longest column of 120 cm but decreased with an increase in E on all three column lengths. Interestingly, the 67.5 and 90 cm columns provided equivalent analysis time. Based on the trade-off between resolution and analysis time, 90 cm column length and 250 V/cm were an excellent compromise for maximum separation in a reasonable time. The Ohm law plot (current vs. voltage) of 4-VPBA coated columns at all three lengths was generated for assessing the joule heating. At equivalent voltage, the amount of current generated is highest and lowest on the shortest and longest column, respectively. The linearity (R2) of the plots showed excellent control of joule heating up to 18, 22, and 30 kV on VBTA columns at 67.5, 90, and 120 cm, respectively (Figure 4C).
FIGURE 4.
Effect of column length at constant electric field strength on resolution (A) and analysis time (B). The Ohm law plot (C) shows good to-excellent linearity for different column lengths for the 4-vinylphenylboronic acid (4-VPBA)-modified column. Experimental conditions are the same as in this figure except for background electrolyte (BGE) pH 8.0 (adjusted with 17.4 M HOAc).
3.4 |. Quantitation of endogenous KP metabolites
Because the analyte-free matrix was not available for preparing the calibration curve, we used the standard addition method in combination with the IS method to quantitate endogenous analytes in a healthy human plasma [18, 23]. Isotopic labeled TRP-D8 was used as IS for TRP, whereas tyrosine was used as IS for KYN (as isotopic labeled KYN was unavailable). The quantitation of 3-OHKYN was not pursued because we could not find an appropriate IS when the experiments were conducted. The plasma samples were prepared by SLE using phospholipid removal cartridge. The conventional protein precipitation did not provide efficient cleanup of phospholipids causing unstable retention time and ion suppression. However, the use of SLE allowed efficient protein and phospholipids removal and better retention stability for TRP and KYN. The calibration curves were prepared by plotting the peak area ratio of standard to IS against the added concentrations of standard analytes.
3.4.1 |. Method validation
According to FDA guidelines, the quantitation method for analyzing plasma samples was validated for linearity, sensitivity, precision, accuracy, and recovery [24].
Linearity and sensitivity
The measures of linearity (linear range, linear regression equation, and R2) and sensitivity (limit of detection [LOD] and limit of quantitation [LOQ]) for TRP and KYN are listed in Table 1. The calibration curves prepared on 4 days (data shown for 3 days) for TRP and KYN showed good repeatability as demonstrated by a small standard deviation in the slope value of the linear regression equations. The linearity of the calibration curves was determined from R2 values and above 0.99 indicated good linearity. The R2 values obtained from calibration for three different days were in the range of 0.9978–0.9997 for TRP and 0.9967–9998 for KYN, which showed good linearity. The representative calibration curves of TRP and KYN from day 1 are shown in Figure S3. The sensitivity of the method was determined from the LOD and LOQ. The calculated LOD (at S/N of 3) and LOQ (at S/N of 10) were 27.4–24 and 80–91 ng/mL, respectively, for TRP and KYN.
TABLE 1.
Linearity, regression equations, limit of detection (LOD), limit of quantitation (LOQ), precision, and accuracy were obtained for tryptophan (TRP) and kynurenine (KYN).
| Tryptophan | Kynurenine | |
|---|---|---|
| Linear range (μg/mL) | 3.8–230 | 0.23–13.3 |
| Calibration curve (n = 3) | y = 8.8(±0.7)x – 100.2(±2) | y = 456(±14)x – 5.2(±0.1) |
| R 2 | 0.9978–0.9997 | 0.9967–0.9998 |
| LOD (ng/mL) | 27.4 | 23.9 |
| LOQ (ng/mL) | 91.1 | 80.2 |
| Intraday %RSD of peak area (n = 3) | 5 | 0.2 |
| 2 | 5 | |
| 1 | 2 | |
| Intraday %RSD of retention time (n = 15) | 3 | 3 |
| Intraday %RSD of peak area (n = 12) | 14* | 9 |
| 5 | 5 | |
| 6 | 17* | |
| Inter-day %RSD of retention time (n = 45) | 10 | 10 |
| Inter-day %RSD of peak area (n = 9) | 17* | 10 |
| 6 | 4 | |
| 8 | 18* |
indficate the highest % RSD for peak area observed among the three spike levels.
Precision, accuracy, and recovery
The intraday and inter-day precision of peak areas of TRP and KYN in spiked plasma samples determined from %RSD is shown in Table S3. The precision of the peak area ratio (Std./IS) was determined at low, medium, and high calibration points. The intraday precision of the peak area ratio of both TRP and KYN ranged from 0.2% to 5% (n = 3). The inter-day precisions of peak area ratios for TRP and KYN were ≤20% RSD at all three levels.
The accuracy of spiking, the precision of quantitation, and the extraction recoveries were assessed at low, medium, and high spiked concentrations of TYP and KYN over 4 days (Table S3). The accuracy of the spiking of plasma samples was calculated from the percent deviation of measured spiked concentration from endogenous concentration plus the added concentration of analytes. The accuracy of spiking on four different days was found to be ≤10% for both TRP and KYN at all three spiked levels except for KYN at a medium spike level on day 2 (Table S3, column 6, row 2, 18% RSD) and low spike level on day 4 (Table S3, column 5, row 4, 15% RSD). The precision of quantitation was determined from the %RSD of measured spiked concentrations. The precision of spiking on four different days was found to be ≤10% for both TRP and KYN at all three spiked levels except for KYN at a medium spike level on day 2 (Table S3, column 6, row 2, 18% RSD) and low spike level on day 4 (Table 2, column 5, row 4, 15% RSD). The inter-day spiking was ≤10% at all three spiked levels for TRP and KYN. The intraday and inter-day spiking precision was 1%–7% RSD for TRP. The intraday precision of the spike for KYN was 2%–7% RSD except at a medium spike level on day 3 (Table 2, column 6, row 8, 14% RSD). The inter-day precision of spiking for KYN was 18% and 13% RSD for low and high spike levels, respectively, whereas it was found to be 5% RSD for high spike levels.
TABLE 2.
Plasma level of tryptophan (TRP), kynurenine (KYN), and TRP/KYN ratio determined in commercial plasma.
| TRP | KYN | KYN/TRP | |||
|---|---|---|---|---|---|
| μM | μg/mL | μM | μg/mL | μM/μM | |
| Day 1 | 99 | 20 | 5.1 | 1.07 | 0.052 |
| Day 2 | 103 | 21 | 5.4 | 1.12 | 0.052 |
| Day 3 | 99 | 20 | 5.1 | 1.06 | 0.052 |
| Day 4 | 90 | 18 | 5.2 | 1.08 | 0.058 |
| Avg. | 98 | 20 | 5.2 | 1.08 | 0.053 |
| SD | 5 | 1 | 0.1 | 0.03 | 0.003 |
| %RSD | 6 | 6 | 2 | 2 | 5 |
The recovery of the spike was calculated from the difference between measured spiked concentrations and measured endogenous concentrations divided by added analyte concentration. The extraction recoveries for TRP on four different days were 77%–126%, 93%–113%, and 92%–106% for low, medium, and high spiked levels, respectively, whereas the inter-day recoveries were 78%–106%, 95%–107%, and 95%–103% for low, medium, and high spiked levels, respectively. Similarly, the extraction recoveries for KYN on four different days were 74%–115%, 82%–136%, and 86%–106% for low, medium, and high spiked levels, whereas inter-day recoveries were 73%–109%, 96%–124%, and 94%–104% for low, medium, and high spiked concentrations, respectively. However, the extraction recovery of KYN at a low spiked level was found to be quite low (i.e., 66%) on day 4.
Endogenous concentrations of TRP and KYN were measured in one healthy volunteer’s plasma sample on four different days using separate calibration curves generated each day. The results of concentrations of TRP and KYN found on four different days are shown in Table 2. The mean endogenous concentration of TRP was 98 μM (±5 μM), and that of KYN was 5.2 μM (±0.1 μM). Our result aligned with previous studies on TRP (45–95 μM) and KYN (1.1–5.9 μM) for the normal concentration levels in adult human blood reported in the human metabolome database [25, 26]. The representative chromatogram of endogenous TRP and KYN in one healthy volunteer plasma sample spiked with IS is shown in Figure 5A. Although TRP and KYN are well resolved, the respective ISs TRP-D8 and TYR are coeluted. However, the EIC (Figure 5B) displays adequate selectivity due to different MRM transitions that were used for quantitation without interference from any coeluting metabolites.
FIGURE 5.
Representative chromatogram of endogenous tryptophan (TRP) and kynurenine (KYN) in healthy human plasma spiked with internal standard (IS) (12 μM TRP-D8 and TYR). (A) Total ion chromatogram, (B) extracted ion chromatograms. Experimental conditions: column: 90 cm, 50 μm i.d., 16 mg/mL 4-vinylphenylboronic acid (4-VPBA) coated, background electrolyte (BGE): 15 mM NH4CO3, pH 8 (adjusted by HOAc), injection: 5 mbar, 20 s, voltage: +25 kV.
A total of 34 plasma samples collected from 17 subjects at times t1 and t2 were used to measure TRP and KYN levels using the developed quantitation method. These patients were identified in low and high neuroinflammation groups based on C-reactive protein (CRP) levels and TNF-α. For example, patients with CRP levels >3 mg/L are identified as high inflammation patients and CRP levels <3 mg/L are identified as low inflammation patients. The mean concentrations of TRP and KYN in low and high neuroinflammation patient’s plasma collected at t1 and t2 are listed in Table S4. Our result indicated no significant difference in average endogenous TRP and KYN concentrations in human subjects with low versus high neuroinflammation. The average KYN/TRP ratio was also not statistically different in subjects with low and high neuroinflammation. These results suggest that sizeable populations must be screened to provide a detailed comparison of human subjects’ metabolite and metabolite ratio between low and high inflammation groups.
4 |. CONCLUDING REMARKS
Using 4-VPBA as a novel bonded stationary phase, we evaluated the concentration of VPBA monomer, mobile phase pH, and concentration of BGE, as well as the column lengths and field strengths to optimize the separation of TRP, KYN, and 3-OHKYN. Next, we developed a sensitive and repeatable quantitative CEC–MS method for measuring endogenous TRP and KYN using combined standard addition and internal method. The excellent stability of this VBTA CEC–MS column at pH 8 and excellent control of joule heating at high voltage with high repeatability of retention times demonstrated the usefulness of CEC–MS for clinical validation and biomarker discovery. The unique combination of liquid–liquid along with solid phase extraction for protein and phospholipid cleanup of human plasma samples was demonstrated in this study. The developed quantitation method was successfully validated and applied for measuring endogenous levels of TRP and KYN in healthy and neuroinflammation human subjects’ plasma samples.
Supplementary Material
ACKNOWLEDGMENTS
The authors gratefully acknowledge the financial support for this project by the National Institutes of Health (1 R21 MH107985-01).
Abbreviations:
- 3-OHKYN
3-hydroxykynurenine
- 4-VPBA
4-vinylphenylboronic acid
- 5-OHTRP
5-hydroxytryptophan
- CRP
C-reactive protein
- FS
fused silica
- IDO
indoleamine 2,3-dioxygenase
- IS
internal standard
- KP
kynurenine pathway
- KYN
kynurenine
- MRM
multiple reaction monitoring
- SLE
solid–liquid extraction
- TDO
tryptophan 2,3-dioxygenase
- TNF-α
tumor necrosis factor-α
- TRP
tryptophan
- ϒ-MAPS
[ϒ-(methacryloyloxy)propyl]trimethoxysilane
Footnotes
CONFLICT OF INTEREST
The authors have declared no conflict of interest.
SUPPORTING INFORMATION
Additional supporting information can be found online in the Supporting Information section at the end of this article.
DATA AVAILABILITY STATEMENT
The data that support the finding of this research work are available in the Supporting Information section of this article.
REFERENCES
- 1.Chen Y, Guillemin GJ. Kynurenine pathway metabolites in humans: disease and healthy States. Int J Tryptophan Res. 2009;2:1–19. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Richard DM, Dawes MA, Mathias CW, Acheson A, Hill-Kapturczak N, Dougherty DM. L-Tryptophan: basic metabolic functions, behavioral research and therapeutic indications. Int J Tryptophan Res. 2009;2:45–60. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Badawy AA, Dougherty DM. Assessment of the human kynurenine pathway: comparisons and clinical implications of ethnic and gender differences in plasma tryptophan, kynurenine metabolites, and enzyme expressions at baseline and after acute tryptophan loading and depletion. Int J Tryptophan Res. 2016;9:31–49. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.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(1):37–44. [DOI] [PubMed] [Google Scholar]
- 5.Möller M, Du Preez JL, Harvey BH. Development and validation of a single analytical method for the determination of tryptophan, and its kynurenine metabolites in rat plasma. J Chromatogr B: Anal Technol Biomed Life Sci. 2012;898:121–9. [DOI] [PubMed] [Google Scholar]
- 6.Orsatti L, Speziale R, Orsale MV, Caretti F, Veneziano M, Zini M, et al. A single-run liquid chromatography mass spectrometry method to quantify neuroactive kynurenine pathway metabolites in rat plasma. J Pharm Biomed Anal. 2015;107:426–31. [DOI] [PubMed] [Google Scholar]
- 7.Fuertig R, Ceci A, Camus SM, Bezard E, Luippold AH, Hengerer B. LC-MS/MS-based quantification of kynurenine metabolites, tryptophan, monoamines and neopterin in plasma, cerebrospinal fluid and brain. Bioanalysis. 2016;8(18):1903–17. [DOI] [PubMed] [Google Scholar]
- 8.Christmas DM, Potokar J, Davies SJ. A biological pathway linking inflammation and depression: activation of indoleamine 2,3-dioxygenase. Neuropsychiatr Dis Treat. 2011;7:431–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Myint AM, Kim YK. Network beyond IDO in psychiatric disorders: revisiting neurodegeneration hypothesis. Prog Neuropsychopharmacol Biol Psychiatry. 2014;48:304–13. [DOI] [PubMed] [Google Scholar]
- 10.Zuo H, Ueland PM, Ulvik A, Eussen SJ, Vollset SE, Nygård O, et al. Plasma biomarkers of inflammation, the kynurenine pathway, and risks of all-cause, cancer, and cardiovascular disease mortality: the Hordaland Health Study. Am J Epidemiol. 2016;183(4):249–58. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Arnhard K, Pitterl F, Sperner-Unterweger B, Fuchs D, Koal T, Oberacher H. A validated liquid chromatography-high resolution-tandem mass spectrometry method for the simultaneous quantitation of tryptophan, kynurenine, kynurenic acid, and quinolinic acid in human plasm. Electrophoresis. 2018;39:1171–80. [DOI] [PubMed] [Google Scholar]
- 12.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. [DOI] [PubMed] [Google Scholar]
- 13.Huang Y, Louie A, Yang Q, Massenkoff N, Xu C, Hunt PW, et al. A simple LC-MS/MS method for determination of kynurenine and tryptophan concentrations in human plasma from HIV-infected patients. Bioanalysis. 2013;5(11):1397–407. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Eser B, Özkan Y, Sepici Dinçel A. Determination of tryptophan and kynurenine by LC-MS/MS by using amlodipine as an internal standard. J Am Soc Mass Spectrom. 2020;31(2):379–85. [DOI] [PubMed] [Google Scholar]
- 15.Notarangelo FM, Wu HQ, Macherone A, Graham DR, Schwarcz R. Gas chromatography/tandem mass spectrometry detection of extracellular kynurenine and related metabolites in normal and lesioned rat brain. Anal Biochem. 2012;421(2):573–81. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Arvidsson B, Johannesson N, Citterio A, Righetti PG, Bergquist J. High throughput analysis of tryptophan metabolites in a complex matrix using capillary electrophoresis coupled to time-of-flight mass spectrometry. J Chromatogr A. 2007;1159(1–2):154–8. [DOI] [PubMed] [Google Scholar]
- 17.Wang X, Davis I, Liu A, Miller A, Shamsi SA. Improved separation and detection of picolinic acid and quinolinic acid by capillary electrophoresis-mass spectrometry: application to analysis of human cerebrospinal fluid. J Chromatogr A. 2013;1316:147–53. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Chawdhury ASMMR, Shamsi SA, Miller A, Liu A. Capillary electrochromatography-mass spectrometry of kynurenine pathway metabolites. J Chromatogr A. 2021;1651:462294. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Rossetti C, Levernæs MC, Reubsaet L, Halvorsen TG. Evaluation of affinity-based serum clean-up in mass spectrometric analysis: plastic vs monoclonal antibodies. J Chromatogr A. 2016;1471:19–26. [DOI] [PubMed] [Google Scholar]
- 20.El Deeb S, Wätzig H, Abd El-Hady D, Sänger-van de Griend C, Scriba GK. Recent advances in capillary electrophoretic migration techniques for pharmaceutical analysis (2013–2015). Electrophoresis. 2016;37(12):1591–608. [DOI] [PubMed] [Google Scholar]
- 21.Liu Y, Shamsi SA. Chiral capillary electrophoresis-mass spectrometry: developments and applications in the period 2010–2015: a review. J Chromatogr Sci. 2016;54(10):1771–86. [DOI] [PubMed] [Google Scholar]
- 22.Neville D, Houghton R, Garrett S. Efficacy of plasma phospholipid removal during sample preparation and subsequent retention under typical UHPLC conditions. Bioanalysis. 2012;4(7):795–807. [DOI] [PubMed] [Google Scholar]
- 23.Hewavitharana AK, Abu Kassim NS, Shaw PN. Standard addition with internal standardisation as an alternative to using stable isotope labelled internal standards to correct for matrix effects—comparison and validation using liquid chromatography-tandem mass spectrometric assay of vitamin D. J Chromatogr A. 2018;1553:101–7. [DOI] [PubMed] [Google Scholar]
- 24.Bioanalytical method validation: guidance for industry. US Health and Human Resources, Food and Drug Administration, Center for Veterinary Medicine, Biopharmaceutics, 2018, Rockville, MD. https://www.fda.gov/media/70858/download [Google Scholar]
- 25.The Human Metabolome Database. https://hmdb.ca/metabolites/HMDB0000929
- 26.The Human Metabolome Database. https://hmdb.ca/metabolites/HMDB0000684
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The data that support the finding of this research work are available in the Supporting Information section of this article.