Tryptophan, an essential amino acid, and its metabolites are involved in many physiological processes including neuronal functions, immune system, and gut homeostasis. Alterations to tryptophan metabolism are associated with various pathologies such as neurologic, psychiatric disorders, inflammatory bowel diseases (IBD),
Trang 1Contents lists available at ScienceDirect
journal homepage: www.elsevier.com/locate/chroma
Aurore Desmonsa, ∗, Lydie Humberta, Thibaut Eguethera, Pranvera Krasniqia,
Dominique Rainteaua, Tarek Mahdib, Nathalie Kapelb, Antonin Lamazièrea
a Clinical metabolomic department, Sorbonne Université, INSERM, Centre de Recherche Saint-Antoine (CRSA), Saint Antoine Hospital, Assistance Publique des
Hôpitaux de Paris (AP-HP.Sorbonne Université), Paris, France
b Laboratoire de Coprologie Fonctionnelle, Hôpitaux Universitaires Pitié-Salpêtrière - Charles Foix, Assistance Publique des Hôpitaux de Paris
(AP-HP.Sorbonne Université), Paris, France
Article history:
Received 12 July 2022
Revised 12 October 2022
Accepted 23 October 2022
Available online 3 November 2022
Keywords:
tryptophan metabolites profile
LC-MS/MS
Inflammatory bowel diseases
Tryptophan,anessentialaminoacid,anditsmetabolitesareinvolvedinmanyphysiologicalprocesses in-cludingneuronalfunctions,immunesystem,andguthomeostasis.Alterationstotryptophanmetabolism areassociatedwithvariouspathologiessuchasneurologic,psychiatricdisorders,inflammatorybowel dis-eases(IBD),metabolicdisorders,andcancer.Itisconsequentlycriticaltodevelopareliable,quantitative methodfortheanalysisoftryptophananditsdownstreammetabolitesfromthekynurenine,serotonin, andindolespathways.AnLC-MS/MSmethodwasdesignedfortheanalysisoftryptophanand20ofits metabolites,withoutderivatizationandperformedinasinglerun.Thismethodwasvalidatedforboth serumandstool.Thecomparisonsbetweenserumandplasma,collectedwithseveraldiffering anticoag-ulants,showedsignificantdifferencesonlyforserotonin.Referencesvalueswereestablishedinseraand stoolsfromhealthydonors.Forstoolsamples,asaproofofconcept,thedevelopedmethodwasapplied
toahealthycontrolgroupandanIBDpatientgroup.Resultsshowed significantdifferencesinthe con-centrationsoftryptophan,xanthurenicacid,kynurenicacid,indole-3-lacticacid,andpicolinicacid.This methodallowedanextensiveanalysisofthethreetryptophanmetabolicpathwaysintwocompartments BeyondtheapplicationtoIBDpatients,theclinicaluseofthismethodiswide-rangingandmaybe ap-pliedtootherpathologicalconditionsinvolvingtryptophanmetabolism,suchasneurological,psychiatric,
orauto-inflammatorypathologies
© 2022 The Authors Published by Elsevier B.V ThisisanopenaccessarticleundertheCCBY-NC-NDlicense (http://creativecommons.org/licenses/by-nc-nd/4.0/)
1 Introduction
Tryptophan (Trp) is an essential aromatic amino acid involved
in protein synthesis and is the precursor of many bioactive com-
pounds Trp and its metabolites are implicated in many physiolog-
ical processes such as neuronal functions, immunoregulation, and
inflammation Trp is also identified as a key marker in gut home-
ostasis and its metabolism is closely linked to the intestinal micro-
biome Around 90% of Trp is metabolized through the kynurenine
✩ Disclosure statement: no
✩✩ All authors declare no competing financial interests and consent for publication
★ Data and material are available
∗ Corresponding author at: Hôpital Saint-Antoine, AP-HP Sorbonne Université, 27,
rue Chaligny, 75012 Paris, France
E-mail address: aurore.desmons@aphp.fr (A Desmons)
pathway, also called the indoleamine-2,3-dioxygenase (IDO) path- way [1–3] This pathway leads to kynurenine production as well
as other neurologically active compounds such as kynurenic acid (KA), quinolinic acid (QA) and 3-hydroxykynurenine (3-HK) QA and 3-HK have neurotoxic properties while KA has neuroprotective effects [4] The second pathway utilizing Trp, which is quantita- tively less important and contributes less than 5% of Trp degrada- tion, is the serotonin pathway This pathway leads to the produc- tion of 80% of total serotonin by intestinal compartment and plays
an important role in neurotransmission and neurological functions [ 3, 5] The last pathway, the aryl hydrocarbon receptor (AhR) path- way utilizes Trp in the synthesis of indole and indoles derivatives
by intestinal bacteria Many of these derivatives, such as indole- 3-acetic acid (IAA), indole-3-propionic acid (IPA), and indole-3-
https://doi.org/10.1016/j.chroma.2022.463602
0021-9673/© 2022 The Authors Published by Elsevier B.V This is an open access article under the CC BY-NC-ND license ( http://creativecommons.org/licenses/by-nc-nd/4.0/ )
Trang 2Fig 1 Tryptophan metabolism via kynurenine (blue), serotonin (green) and indoles (orange) pathways
Chemicals structures of tryptophan and metabolites quantified by the LC-MS/MS method developed Figure adapted from Agus et al [3]
carboxaldehyde, can activate the AhR expressed by some immune
and intestinal cells [ 6, 7]
Alterations to Trp metabolism are associated with many patho-
logical states such as neurodevelopmental, neurologic and psy-
chiatric disorders, metabolic disorders, and cancer [ 8, 9] Many
publications have highlighted the modifications of tryptophan
metabolism in gastrointestinal disorders including inflammatory
bowel diseases (IBD) and irritable bowel syndrome (IBS) [ 3, 10] In
IBD, these alterations of Trp metabolism may be involved in the
pathogenesis of the disease [ 3, 11] Indoles metabolites are highly of
interest in several disorders, a recent study reported that metabolic
disorders are associated with a decrease of AhR agonists produced
from trp [12] Depletion of trp metabolites including AhR agonists
may be affect the severity of the disease [13] It is consequently
critical to develop a reliable, quantitative method for the analysis
of Trp and its metabolites for research and clinical purposes
To explore Trp metabolism, several quantitative methods have
been developed The first methods consisted of liquid chromatog-
raphy (LC) separation associated with detection based on UV ab-
sorbance, fluorescence, or electrochemistry [14] Recent methods
have been developed using LC coupled to tandem mass spectrom-
etry (LC-MS/MS) focusing on the major metabolites of kynurenine
and serotonin pathways, and quantification performed mainly on
serum or plasma matrices [ 15, 16] Global analysis of Trp metabo-
lites was developed in different biological samples but quantifica-
tions were dedicated to non-human matrices [ 17, 18] The purpose
of our study was to explore the metabolism of Trp in the two com-
partments (i.e matrices) relevant in IBD and IBS contexts [ 12, 13]
Beyond theses pathologies, exploration of trp metabolism is highly
of interest in neurological diseases, such as multiple sclerosis and
Huntington’s disease [ 19, 20] This panel may be useful to follow
new therapies targeting the gut microbial metabolism [13] Pub-
lished methodology studies have mainly been performed either in
serum or stool, but not in both compartments of patients
Here, we developed a LC-MS/MS method for the quantification
of tryptophan and 20 of its metabolites in three different human
biological matrices: feces, serum, and plasma ( Fig.1) This method
allows an extensive analysis of the three stated metabolic path-
ways in a single test, without derivatization Quantification was es-
tablished and validated in human serum, plasma, and stool using both internal standards and external calibration, which allowed the establishment of references values in the two compartments As a proof of concept, the developed method was applied to a control healthy group and a patients group diagnosed with IBD, including ulcerative colitis (UC) and Crohn’s disease
2 Material and methods
2.1 Chemicals and reagents
Tryptophan (Trp) ( ≥ 99 %), picolinic acid (Pico) (99 %), quino- linic acid (QA) (99 %), 3-hydroxy-kynurenine (3HK) ( ≥ 98 %), serotonin (5HT) ( ≥ 98 %), 5-hydroxy-tryptophan (5HTP)( ≥ 98 %), 3-hydroxy-anthranilic acid (3-HAA) (97 %), kynurenine (KYN) ( ≥
98 %), xanthurenic acid (XA) (96 %), tryptamine (TA) ( ≥ 97 %), kynurenic acid (KA) ( ≥ 98 %), 5-hydroxyindole acetic acid (5- HIAA) ( ≥ 98 %), N-acetylserotonin (NAS) ( ≥ 99 %), indole-3- acetamide (IAM) (98 %), indole-3- lactic acid (ILA) (99 %), indole- 3-carboxaldehyde (I-3CA) (97 %), melatonin (MELA) ( ≥ 98 %), tryp- tophol (TOL) (97 %), indole-3-acetic acid (IAA) ( ≥ 98 %), indole- 3-propionic acid (3-IPA) (99 %), indoxyl-sulfate (3-IS) ( ≥ 97 %) and methanol (hypergrade for LC-MS LiChrosolv®) were purchased from Sigma-Aldrich (St Louis, Missouri, USA) Chemicals purity were provided in brackets for each standards molecules
Internal standards (ISs), L-Tryptophan-d8 (Trp-d8) (98 %) and anthranilic acid 15N (A15N) (98 %) were obtained from Eurisotop (Saint-Aubin, France) and from Cambridge Isotope Laboratories, Inc (Andover,USA), respectively
Formic acid (FA) was acquired from Honeywell-Fluka Fisher scientific (IllKirch France), HiPerSolv CHROMANORM acetonitrile (ACN) for HPLC LC-MS grade from VWR (Radnor, USA)
2.2 Analytical protocol 2.2.1 Calibration standards
The stocks solutions of unlabeled standards (1mg/mL for each molecule) were prepared in water/methanol (50/50) (v/v) for KYN and I3CA TA, TOL, IAM, IAA, 3-IPA, 5-HIAA, ILA, MELA,
Trang 33-IS, Pico and NAS were dissolved in methanol (1mg/mL for
each molecule) XA, KA were dissolved in dimethylsulfoxide
(DMSO) at 1mg/ml QA, 3-HAA were dissolved in dimethylsulfox-
ide (DMSO)/methanol 10/90 (v/v) at 1mg/ml 5HT, 5HTP, were dis-
solved in water/methanol/acetic acid 90/9.9/0.1 (v/v/v) at 1mg/ml
3HK was dissolved in water/methanol/NaOH0.2M 90/9.9/0.1 (v/v/v)
at 1mg/ml Trp solution (2mg/ml) was prepared in water/sodium
hydroxide 0,2 M (75/25) Stock solutions of ISs Trp-d8 (1mg/ml)
and A15N (1mg/ml) were dissolved in methanol
For serum, a stock calibration standards solution was prepared
by mixing equal volumes of the twenty tryptophan metabolites
Twelve calibration standards levels were prepared from the stock
solution (ranging from 23.8 μg/ ml to 3 ng/ml) by serial dilutions
in methanol Seven levels of calibration standards were prepared
separately for tryptophan (ranging from 125 μg/ml to 60 ng/ml) by
serial dilutions in methanol
All stocks solutions were stored at -80 °C until analysis
2.2.2 Serum calibration standards preparation
20 μL of calibration standards were transferred to 1.5 mL Ep-
pendorf tube, 50 ng of each internal were added from stock so-
lution (1mg/mL) for Trp d-8 and A15N, 50 μL of “lyophilized
blank serum” (Chromsystems, Gräfelfing, Germany) and 430 μL of
methanol were added
Samples were mixed for 15 and placed for 30 min at 4 °C After
centrifugation (18 0 0 0 g for 15 min at +4 °C), the supernatant was
transferred to a glass vial We injected 2 μl of the mix into the
HPLC-MS/MS system for analysis
2.2.3 Serum and plasma samples preparation
50 μL of plasma or serum sample were transferred to 1.5 mL
Eppendorf tube, 50 ng of each internal were added from stock
solution (1mg/mL) for Trp d-8 and A15N, and 450 μL of cold
methanol was added The samples were vortex-mixed for 15 and
placed for 30 min at 4 °C The samples were centrifuged (18 0 0 0
g for 15 min at +4 °C) The total content of the supernatant was
transferred to a vial from which 2 μl were injected into the HPLC-
MS/MS system for analysis
2.2.4 Stool calibration standards preparation
20 μL of calibration standards were transferred to 1.5 mL Ep-
pendorf tube, 50 ng of each internal were added from stock solu-
tion (1mg/mL) for Trp d-8 and A15N, and 480 μL of methanol were
added Samples were vortex-mixed again for 15 and transferred
into glass vials before injection of 2 μL into the HPLC-MS/MS sys-
tem for analysis
2.2.5 Stool samples preparation
Stool samples were frozen at -80 °C then, all samples were
feeeze-dried (Freeze dryer Buchi L-200 with scroll pump 6 qm/h)
at -55.9 °C, 0.300 mbar, for 24h 10 mg of freeze-dried stool were
transferred to 1.5 mL Eppendorf tube and 500 μL of milliQ water
were added The samples were vortex-mixed for 15 s 500 μL of
methanol and 50 ng of each internal were added from stock so-
lution (1mg/mL) for Trp d-8 and A15N and samples vortex-mixed
again for 15 s After centrifugation at 18 0 0 0 g for 15 min at + 4 °C,
the supernatant is collected A second extraction was made by
adding 1 ml of methanol / water (90/10) (v/v) to the pellet Sam-
ples were vortex-mixed for 15 s After centrifugation (18 0 0 0 g for
15 min at +4 °C) the two supernatants were pooled Evaporation
was performed under a nitrogen flow Five hundred μl of methanol
were added to the residue Five μl were injected onto the HPLC-
MS/MS system for analysis
2.3 HPLC-ESI-MS/MS analysis
Samples were analyzed using a LC-20ADXR (Shimadzu, Kyoto, Japan) chromatographic system in tandem with a linear ion trap quadrupole MS/MS spectrometer QTRAP 5500 system (SCIEX, On- tario, Canada) Chromatographic separation was performed with a kinetex biphenyl column (100 ×2.1 mm; particle size 2.6μm) (Phe- nomenex, Torrance, USA) with a 2.1 mm C8 SecurityGuard TM UL- TRA Cartridges UHPLC guard column (Phenomenex, Torrance, USA) The mobile phases were composed of 0.4 % formic acid (FA) in water (v/v) (mobile Phase A) and 0.4% FA in acetonitrile (mobile phase B) The column temperature was set at 17 °C, the gradient elution was performed at 0.3 mL/min, starting at 3% of phase B; then increasing to 20 % of phase B from 0 to 0.1 min; then increas- ing to 65 % B from 0.1 to 7 min; finally increasing to 95 % of phase
B from 7 to 7.5 min, and completed to 2.5 min at 95% of phase B for wash The column reequilibration consisted in a plateau of 3%
B for 2.5 min at 0.3 mL/min The injection volume was 2 μL (for plasma, serum, standards and stool)
MS detection was performed using electrospray ionisation (ESI)
in positive and negative modes, using the Multiple Reaction Moni- torig (MRM) function of the analyser For each analyte, the MS con- ditions were determined via direct infusion of individual standard solutions.”
Compressed air was used as the desolvation gas and nitrogen was used as the collision gas The instrument parameters were set
as follows: nebuliser gas and turbo gas: 40 psi, curtain gas: 20 psi, ion spray voltage: +/- 4500 V for positive or negative ioniza- tion respectively, source temperature: 450 °C Declustering poten- tials (DPs) were set at 60 V, except for QA and Pico, at 33 V A dwell-time of 10 ms was set for all transitions at positive ioniza- tion and 20 ms for negative ionization
2.4 Data processing
Spectral data acquisitions were processed using Analyst (v1.6.3) software and quantifications were performed with Multiquant (v3.0.2) software (SCIEX, Ontario, Canada) GraphPad Prism (v6.01) (San Diego, CA, USA) and SIMCA (v16.0.2.10561) (Göttingen, Ger- many) softwares were used for statistical analysis
2.5 Method validation
The method validation was based on the recommendations of
NF EN ISO 15189 criteria and international guidelines dedicated to LC-MS/MS methods [21–23]
2.5.1 Linear range
Suitable calibration ranges were determined based upon analy- sis of pooled sample of serum and adjusted accordingly For each analyte using an internal standard, peak area response ratios were calculated and plotted against the nominal concentration A linear
fit was employed, and a 1/x weighting factor was applied
2.5.2 Lower limit of detection (LOD) and lower limit of quantification (LLOQ)
To determine LOD and LLOQ, analysis of spiked samples with decreasing concentrations of analytes were performed The LOD was accepted for a signal-to-noise ratio (S/N) ≥ 3 and the LLOQ for a signal-to-noise ratio (S/N) ≥ 10
2.5.3 Intra- and interday accuracy and precision
Accuracy and precision were estimated from the analysis of 2 levels of quality control (QC) samples prepared for each analyte For intraday analysis, 20 samples were prepared and assayed in
Trang 4the same day for serum and stool For interday analysis, 6 sam-
ples and 3 samples were analyzed in separate days for serum and
stool, respectively Acceptance criteria for accuracy was determined
as a bias within ± 15% of the nominal value and within ± 20% of
the LLOQ Acceptance criteria for precision were defined as within
15% of relative standard deviation (R.S.D) and 20% of R.S.D of the
LLOQ
2.5.4 Stability
Stability was evaluated using serum calibration standards ob-
tained as described in part 2.2.2 Calibration standards were pre-
pared and quantified immediately, which served as a reference
Other aliquots from the same solution were stored at -80 °C These
samples were subjected to three freeze-thaw cycles, 10 days, one
month, and 2 months after initial freezing Stability was acceptable
if the mean concentration was between 85% and 115% of the refer-
ence mean concentration
2.5.5 Type of collection tube
The impact of the type of collection tube was determined by
analyzing serum and plasmas collected in citrate, heparin, and
EDTA anticoagulants from 9 healthy donors at fed state
2.5.6 Matrix effects
Peaks areas were determined in 3 different sets of samples: a
blank matrix sample of stool (set 1), a blank matrix sample of stool
spiked with standard solutions at different levels after extraction
(set 2), and one prepared from blank matrix from the same stool
but spiked just before extraction (set 3) The matrix effect was de-
termined as follows and expressed as a percentage: (set 2-set 1) /
(spiked concentration in methanol) An extraction yield was evalu-
ated as follows and expressed as a percentage: (set 3)/(set 2) [21]
2.6 Biological sample testing
2.6.1 Application of the assay to clinical samples
The assay protocol was applied to clinical samples from a con-
trol group and a group of patients diagnosed with IBD (including
UC and Crohn’s disease)
2.6.2 Sample collection information
Plasma and serum samples were obtained from the French
blood establishment (EFS, Saint Antoine Hospital) Stool samples
were received from the routine workload of the laboratory for
the follow-up of patients The study using stool residues was ap-
proved by the French public health organization (CSP-article L1121-
3, amended by the law n °2011-2012, December 29, 2011-article 5)
Blood, from non-fasting or fasting subjects, was collected in
tubes containing coagulant, a clot activator, or anticoagulant (hep-
arin, sodium citrate, EDTA) After collection, tubes were centrifuged
(3,0 0 0 g for 10 min at 8 °C) and serum or plasma aliquoted in Ep-
pendorf tubes, ensuring that no red blood cells or clots were car-
ried over, and stored at -80 °C before analysis Stool samples were
collected and immediately stored at -80 °C; samples were freeze-
dried before analysis
3 Results and discussion
3.1 LC-MS/MS method development
For a specific detection, MRM transitions were selected for each
metabolite and internal standard, as shown in Table1 A represen-
tative chromatogram of all targeted metabolites was shown in Fig
2 Ion products resulting from the losses of 18 or 44 m/z, corre-
sponding to the loss of stable neutral molecules H 2O and CO 2 re-
spectively, were excluded where possible This allowed us to ob-
tain more specific transitions, and to avoid cross contamination
from these common losses For molecules with very close precur- sor ions (m/z 206.1 for XA and 206.2 for ILA), product ions used for quantification were different (m/z 160.1 and 130.1 for XA and ILA, respectively) and the chromatographic retentions times were also different (RT: XA 2.2 min and ILA 3.3 min)
For KA and 3IPA, with the same precursor ion (m/z 190.1), the assigned product ions were different (m/z 115.9 KA and 130.1 IPA), and peaks were chromatographically resolved (RT: 2.4 min for KA and 4.6 min for 3IPA) The same ion products were shared for quantification of some indoles derivatives (m/z 130.1): for IAM, ILA, 3IPA but the peaks were chromatographically resolved (RT: 3.0 min (IAM); 3.3 min (ILA); 4.6 min (3-IPA)) All metabolites, except for indoxyl sulfate, were analyzed using a positive ESI and were all metabolites were tuned for maximum sensitivity across the linear range ( Table1)
A challenge for the development of this method was the pres- ence of metabolites in serum, plasma, and stool at very differ- ent levels This heterogeneity in metabolite concentrations required
an instrument with a great dynamic range and the appropriate preparation of standard solutions at linear ranges Thus, two solu- tions were prepared, one containing 20 metabolites, ranging from
3 ng/ml to 23.8 μg/ml, and another for Trp, ranging from 60 ng/ml to 125 μg/ml Both solutions were obtained by serial dilu- tions in methanol During preparation of stock solution, poor sol- ubility and stability of some metabolites in methanol or in aque- ous solutions were observed This phenomenon of instability was previously described for some metabolites from the kynurenines pathway in aqueous solutions (16) For these metabolites, a wa- ter/methanol mix, with the addition of sodium hydroxide, acetic acid, or DMSO, was required, as detailed in the materials and methods section For chromatographic separation, a C18 column was tested first (Kinetex® 5 μm, C18, 100 ˚A, 100 ×2.1 mm) C18 columns, which use octaldecylsilane, were widely used in other quantitative methods dedicated to Trp and its metabolites because
of the use of non polar solvents for these methods (e.g water, methanol and acetonitrile)[ 14–16, 18] We also tested a biphenyl column, a new phase, that have short alkyl biphenyl ligands co- valently bound to the silica surface, stable under 100% aqueous conditions and exhibited good reverse phase retention and aro- matic selectivity Eventually, separation was performed on this col- umn because we obtained an enhanced retention for the aromatic derivatives
Different column temperatures were used, from 15 °C to 55 °C [ 18, 24, 25], and was maintained at 17 °C following assays, which corresponded to the lowest value available for our device and en- vironment This temperature value, in comparison with higher val- ues, allowed increased retention times and better separation for the compounds eluted first Different percentages of FA were tested
in mobile phases, ranging from 0 to 0.4%, to check for molecule specific ionization; the percentage of 0.4 % was set The presence
of formic acid (FA) at the highest percentage (i.e 0.4 %) in mo- bile phases allowed a better ionization of molecules, improved the shapes of the peak, and targeted analytes exhibited MS bet- ter responses.The elution gradient was tested with H 2O/Methanol and H 2O/ACN respectively for mobile phases A/B; better shapes of peaks were obtained when ACN was used
3.2 Method validation
The method validation was performed based on different inter- national guidelines including Food and Drug Administration (FDA) and European Medicines Agency (EMA) recommendations [21–23] The results for intraday and interday precision and accuracy are reported in Table 2 and 3 for serum and stool All metabolites demonstrated a CV < 20% for the lowest values and a CV < 15% for intraday and interday precisions For the intraday assay, analytical
Trang 5Table 1
MS conditions and retention times for metabolites and labeled internal standards (ordered by retention time) ∗ quantifier ion
Component Names Retention Time (min) Precursor ion (m/z) Products ions (m/z) Collision energy (V) Internal Standard
Positive Mode [M + H] [M + H]
3-Hydroxyanthranilic acid 2.1 154.1 ∗ ; 136.1 136.1 ∗ ; 108.2 25 ∗ ; 25 L-Tryptophan-d8
Internal Standards
Fig 2 Chromatogram of the 21 standards of analysis ordered by retention times (in brackets, transitions of precursor and quantifier ions (m/z))
accuracy was < 15% (recovery comprised between 85 % and 115%),
and for the interday assay some recoveries were higher than 15%,
particularly for the lowest values in serum Accuracies in serum
ranged between 85 and 115% except for 3 compounds for which
percentages moderately under 85%: NAS (84%), IAA (84%) and 3-
IPA(83%) ( Table 2) For intraday and interday assays in stool, an-
alytical accuracy and CV were < 15% for all molecules quantified
( Table3) Dynamic ranges were determined according to the phys-
iological and pathological values previously described, and results
are reported in Table 4for LOD and LOQ, and Table2 for ULOQ,
which corresponded to the highest calibration standards quanti-
fied
To account for matrix effect for serum, calibration standards were prepared using lyophilized blank serum For stool samples, matrix effects were evaluated as described in material and meth- ods section Matrix effects calculated at 2 levels were in accor- dance with published studies, ranging from 50 % to 125%, but few components exhibited highest matrix effects (34%, 196%, 41 %, 36
%, for picolinic acid, tryptophan, melatonin, N acetyl serotonin, re- spectively) ( Table 5) These resuts were obtained for the lowest level of spike for all of the 4 metabolites, but the levels of spike tested in our method validation were lower than other published methods [18] Recoveries ranging from 62% to 134% ( Table5) De- spite high variance, these results were consistent with those previ-
Trang 6Table 2
Intraday and interday precision in serum
Means, coefficients of variation (CVs) and recoveries were calculated from internal quality controls (number of assays:
20 for intra- and 6 inter-assays precisions, respectively) for two levels
Component Names Nominal concentration (nM) CV(%) Recovery (%) CV(%) Recovery (%)
ously described in a similar matrix [18] Matrix effects could be
minimized by adding a larger number of stable isotope-labeled
analogue to the targeted molecule Some preanalyticals steps based
on on phospholipid removal and interferent proteins may be per-
formed to minimize matrix effect In blood, method validation was
performed in serum samples, but a comparison was done between
blood collected on different anticoagulants (EDTA, heparin, citrate)
No major difference was found regarding the presence and type of
anticoagulant except for with serotonin, where levels were higher
in serum than in plasma (supplemental data, Fig.1) These results
for serotonin were already described [15], however, it was neces-
sary to evaluate the percentage of difference when developing our
method Residue stabilities are shown in supplemental data, Table
1 After the first thawing, at 10 days, 7 compounds (3-HK, XA, KA,
5-HIAA , I3A , ILA , IS) were under acceptable values
3.3 Application of the method: human serum and stool
concentrations
The main objective of our work was the development of a re-
liable method for the exploration of Trp metabolism in two rele-
vant compartments, serum and stool, in cases of IBD and IBS This
method allows extensive metabolite profiling of Trp metabolism
and highlights changes in both host metabolism (serum) and mi- crobial metabolism (stool) [26] Following analytical validation, the method was tested on 24 human serum samples and 22 human stool samples from healthy donors ( Table6) Reference values were
in agreement with previously published reports [26] The few dif- ferences between our data and published results could be ex- plained by the non-fasting state of our donors, which was im- posed by the composition of the available patient cohort Subse- quently, we tested our method on IBD patient samples to investi- gate its potential in clinical practice Samples were obtained from two groups: patients diagnosed with IBD (27 patients) and a con- trol group of non-IBD patients (13 patients) ( Fig 3A) Fecal cal- protectin was used to classify patients into three different groups among IBD patients according to the established cut-offs used in IBD management and diagnosis [27]: 1) Patients with fecal calpro- tectin < 50 μg/g, 2) patients with fecal calprotectin between 50 and 200 μg/g, which corresponds to an intermediate state, and 3) patients with fecal calprotectin > 200 μg/g which corresponds to
an acute phase of the disease Fecal calprotectin was also assayed
in patients from the control group, who all had a calprotectin level
< 50 μg/g
The multivariate analysis identified the most discriminant vari- ables (VIP > 1) to stratify non IBD patients from IBD patients:
Trang 7Table 3
Intraday and interday precision in stool
A Intraday precison in stool B Interday precision in stool Means, coefficients of variation (CVs) and recoveries were calculated from internal quality controls (number of assays: 20 for intra- and 3 inter-assays precisions, respectively) for two levels
Component Names Mean (nmol/g) CV (%) Mean (nmol/g) CV (%) Mean (nmol/g) CV (%)
A
B
Component Names Mean (nmol/g) CV (%) Mean (nmol/g) CV (%) Mean (nmol/g) CV (%)
Fig 3 Analysis of clinical samples
Stool samples from non IBD and IBD patients A Score plot of patients obtained from OPLS-DA B Variable importance in projection (VIP) exhibited significant differences for XA, KA, ILA, calprotectin, trp, Pico XA, KA, ILA, calprotectin and trp are increased (red bar plot) and pico is decreased (blue bar plot) in IBD patients compared to control group
Trang 8Table 4
Limit of detection (LOD) and limit of quantification (LOQ)
LOD was determined by a signal-to-noise ratio (S/N) of 3:1 and LOQ by a signal-to-noise ratio (S/N) of 10:1
Lower limit of Detection (nmol/L) Lower limit of Quantification (nmol/L)
Table 5
Matrix effect and yield extraction in stool at 3 levels (low, medium, high)
Component Names Spiked Concentration μg/ml Matrix effect % Recovery %
Trang 9Table 6
Metabolites concentrations in serum and feces References values
were determined from healthy individuals in serum (n = 24) and stool
(n = 22), results were expressed as mean ± standard deviation (SD)
Serum (nmol/L) Feces (nmol/g)
3-Hydroxy Kynurenine 12 ± 14 < LOQ
5-Hydroxytryptophan < LOQ 170 ± 180
3-Hydroxyanthranilic acid 33 ± 47 1730 ± 1000
5-Hydroxyindole acetic acid 57 ± 49 230 ± 240
Indole-3-acetamide < LOQ < LOQ
Indole-3- lactic acid 823 ± 228 450 ± 450
Indole-3-carboxaldehyde < LOQ 1150 ± 900
Indole-3- acetic acid 1489 ± 475 5360 ± 4760
Indole-3-propionic acid 1337 ± 1039 6500 ± 3380
Indoxyl sulfate 3130 ± 1451 70 ± 320
calprotectin, Trp, and four Trp metabolites, xanthurenic acid (XA),
kynurenic acid (KA), indole-3-lactic acid (ILA) and picolinic acid
(Pico) ( Fig.3B) These results were confirmed by univariate statis-
tical analysis as shown in supplemental data, Fig.2 The results ob-
tained showed that Trp derivatives are important markers for study
in clinical settings with our method
Three patients (numbers 30, 31, and 37) tested very close to
the control group in the score plot Patient 37 was suffering from
Crohn’s disease and presented 3 consecutive normal levels of cal-
protectin over 6 months, which was probably due to a deep remis-
sion Patient 31 also presented Crohn’s disease in deep remission
and did not receive drug treatment Patient 30 was suffering from
UC under treatment and presented two consecutive normal levels
of calprotectin over 6 months, which was probably due to a steady
state of the disease
Most published studies of Trp derivatives in IBD patients have
been performed in serum or plasma samples Trp metabolites, such
as PA and XA, were reported at a lower concentration in serum or
plasma than control groups [ 16, 26] The increase we observed for
Trp in stool samples supports previous results, which showed an
increase in fecal amino acids [26] Trp increase in feces could be
explained by inflammation and mucosal damage in IBD due to a
compromised gut epithelial barrier [ 26, 28] An extensive quantifi-
cation of amino-acid profiles in blood and stool may be useful to
highlight the impaired absorption of all amino acids in the gut in
patients suffering from IBD
Regarding the indoles pathway, which corresponds to indoles
derivatives produced by commensal bacteria, an increase of ILA
in feces was observed Some studies have shown how indoles
molecules derived from bacteria are implicated in intestinal in-
flammation, IPA, an indole molecule, have been reported to be de-
creased in serum [29] This result is consistent with the increase
of ILA found in feces, a downstream product of IPA Other results
showed an increase for two downstream products of the kynure-
nine pathway, KA and XA, and a decrease in PA The increase of
KA and XA, which are derived from kynurenine metabolism by in-
dolemanine 2,3 dioxygenase (IDO), could be supported by the pre-
viously reported increase of IDO activity in IBD patients [30] A de-
crease for PA in feces may be explained by an increase of the pro-
duction of QA, however, in our study, no significance was found for
the increase in QA A decrease in PA was previously reported in the
serum of IBD patients [ 16, 31] but no studies have determined PA levels in feces for these patients
Eventually, despite a rather small number of patients in the cohort, the results obtained through our developed method evi- denced significative differences of tryptophan derivatives patterns between IBD and non IBD patients
4 Conclusions
A HPLC-ESI-MS/MS method has been developed for the analysis
of Trp and 20 metabolites for application to human serum, plasma, and stool samples This method allows for an extensive analysis of the three Trp metabolic pathways in two compartments and was applied to a clinical study of IBD patients Increased concentrations
of Trp, XA, KA, ILA, and a decrease of PA were observed in IBD pa- tients compared to healthy controls using this approach The clini- cal use of this method is wide-ranging and may be applied to other pathological conditions involving Trp metabolism such as neuro- logical, psychiatric, or auto-inflammatory pathologies
Authors contribution
A.D and A.L performed the design of research L.H., A.D and D.R performed experiments and data acquisition A.D., N.K and A.L performed data analysis and interpretation A.D wrote the pa- per All authors critically revised the article All authors approved the final version to be published
Declaration of Competing Interest
The authors declare that they have no known competing finan- cial interests or personal relationships that could have appeared to influence the work reported in this paper
Data Availability
Data will be made available on request
Supplementary materials
Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.chroma.2022.463602
References
[1] PJ Kennedy, JF Cryan, TG Dinan, G Clarke, Kynurenine pathway metabolism and the microbiota-gut-brain axis, Neuropharmacology 112 (2017) 399–412 [2] I Cervenka, LZ Agudelo, JL Ruas, Kynurenines: Tryptophan’s metabolites in ex- ercise, inflammation, and mental health, Science 357 (6349) (2017) eaaf9794 [3] A Agus, J Planchais, H Sokol, Gut microbiota regulation of tryptophan metabolism in health and disease, Cell Host Microbe 23 (6) (2018) 716–724 [4] L Boulet, G Besson, L Van Noolen, P Faure, ECOPHEN Study Group, F Maillot,
et al., Tryptophan metabolism in phenylketonuria: A French adult cohort study,
J Inherit Metab Dis 43 (5) (2020) 944–951 [5] GM Mawe, JM Hoffman, Serotonin signalling in the gut–functions, dysfunc- tions and therapeutic targets, Nat Rev Gastroenterol Hepatol 10 (8) (2013) 473–486
[6] EE Alexeev, JM Lanis, DJ Kao, EL Campbell, CJ Kelly, KD Battista, et al., Micro- biota-derived indole metabolites promote human and murine intestinal home- ostasis through regulation of interleukin-10 receptor, Am J Pathol 188 (5) (2018) 1183–1194
[7] TD Hubbard, IA Murray, GH Perdew, Indole and tryptophan metabolism: en- dogenous and dietary routes to ah receptor activation, Drug Metab Dispos Biol Fate Chem 43 (10) (2015) 1522–1535
[8] M Platten, EAA Nollen, UF Röhrig, F Fallarino, CA Opitz, Tryptophan metabolism as a common therapeutic target in cancer, neurodegeneration and beyond, Nat Rev Drug Discov 18 (5) (2019 May) 379–401
[9] W Roth, K Zadeh, R Vekariya, Y Ge, M Mohamadzadeh, Tryptophan Metabolism and Gut-Brain Homeostasis, Int J Mol Sci 22 (6) (2021) 2973 [10] RL Burr, H Gu, K Cain, D Djukovic, X Zhang, C Han, et al., Tryptophan metabo- lites in irritable bowel syndrome: an overnight time-course study, J Neurogas- troenterol Motil 25 (4) (2019) 551–562
Trang 10[11] M Sun, N Ma, T He, LJ Johnston, Ma X Tryptophan, Trp) modulates gut home-
ostasis via aryl hydrocarbon receptor (AhR), Crit Rev Food Sci Nutr 60 (10)
(2020) 1760–1768
[12] L Xiao, Q Liu, M Luo, L Xiong, Gut microbiota-derived metabolites in irritable
bowel syndrome, Front Cell Infect Microbiol [Internet] (2021) [cited 2022 Apr
19];11 Available from: https://www.frontiersin.org/article/10.3389/fcimb.2021
729346
[13] HM Roager, TR Licht, Microbial tryptophan catabolites in health and disease,
Nat Commun 9 (1) (2018) 3294
[14] I Sadok, K J ˛edruchniewicz, K Rawicz-Pruszy ´nski, M Staniszewska, UHPLC-E-
SI-MS/MS quantification of relevant substrates and metabolites of the kynure-
nine pathway present in serum and peritoneal fluid from gastric cancer pa-
tients-method development and validation, Int J Mol Sci 22 (13) (2021) 6972
[15] L Boulet, P Faure, P Flore, J Montérémal, V Ducros, Simultaneous determina-
tion of tryptophan and 8 metabolites in human plasma by liquid chromatog-
raphy/tandem mass spectrometry, J Chromatogr B 1054 (2017) 36–43
[16] L Whiley, LC Nye, I Grant, N Andreas, KE Chappell, MH Sarafian, et al., Ul-
trahigh-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 91 (8) (2019) 5207–5216
[17] W Hou, D Zhong, P Zhang, Y Li, M Lin, G Liu, et al., A strategy for the tar-
geted metabolomics analysis of 11 gut microbiota-host co-metabolites in rat
serum, urine and feces by ultra high performance liquid chromatography-tan-
dem mass spectrometry, J Chromatogr A 1429 (2016) 207–217
[18] A Lefèvre, S Mavel, L Nadal-Desbarats, L Galineau, S Attucci, D Dufour,
et al., Validation of a global quantitative analysis methodology of tryptophan
metabolites in mice using LC-MS, Talanta 195 (2019) 593–598
[19] V Rothhammer, ID Mascanfroni, L Bunse, MC Takenaka, JE Kenison, L Mayo,
et al., Type I interferons and microbial metabolites of tryptophan modulate
astrocyte activity and central nervous system inflammation via the aryl hydro-
carbon receptor, Nat Med 22 (6) (2016) 586–597
[20] J Kaye, V Piryatinsky, T Birnberg, T Hingaly, E Raymond, R Kashi, et al.,
Laquinimod arrests experimental autoimmune encephalomyelitis by activating
the aryl hydrocarbon receptor, Proc Natl Acad Sci U S A 113 (41) (2016)
E6145–E6152
[21] FT Peters, OH Drummer, F Musshoff, Validation of new methods, Forensic Sci Int 165 (2–3) (2007) 216–224
[22] A Mochizuki, K Ieki, H Kamimori, A Nagao, K Nakai, A Nakayama, et al., Proposal for risk-based scientific approach on full and partial validation for general changes in bioanalytical method, Bioanalysis 10 (8) (2018) 577–
586 [23] H Blume, E Brendel, M Brudny-Klöppel, S Grebe, B Lausecker, G Rohde, et al., Workshop/conference report on EMA draft guideline on validation of bioana- lytical methods, Eur J Pharm Sci 42 (3) (2011) 300–305
[24] L Schwieler, A Trepci, S Krzyzanowski, S Hermansson, M Granqvist, F Piehl,
et al., A novel, robust method for quantification of multiple kynurenine path- way metabolites in the cerebrospinal fluid, Bioanalysis 12 (6) (2020) 379–
392 [25] Z Galla, C Rajda, G Rácz, N Grecsó, Á Baráth, L Vécsei, et al., Simultaneous determination of 30 neurologically and metabolically important molecules:
A sensitive and selective way to measure tyrosine and tryptophan pathway metabolites and other biomarkers in human serum and cerebrospinal fluid, J Chromatogr A 1635 (2021) 461775
[26] L Aldars-García, JP Gisbert, M Chaparro, Metabolomics insights into inflam- matory bowel disease: a comprehensive review, Pharm Basel Switz 14 (11) (2021) 1190
[27] A Dhaliwal, Z Zeino, C Tomkins, M Cheung, C Nwokolo, S Smith, et al., Utility of faecal calprotectin in inflammatory bowel disease (IBD): what cut-offs should
we apply? Frontl Gastroenterol 6 (1) (2015) 14–19 [28] A Murgia, C Hinz, S Liggi, J Denes, Z Hall, J West, et al., Italian cohort of pa- tients affected by inflammatory bowel disease is characterised by variation in glycerophospholipid, free fatty acids and amino acid levels, Metabolomics 14 (10) (2018) 140
[29] Y Lai, J Xue, CW Liu, B Gao, L Chi, P Tu, et al., Serum metabolomics identifies altered bioenergetics, signaling cascades in parallel with exposome markers in Crohn’s disease, Molecules 24 (3) (2019) 449
[30] MA Ciorba, Indoleamine 2,3 dioxygenase (IDO) in intestinal disease, Curr Opin Gastroenterol 29 (2) (2013) 146–152
[31] LM Chen, CH Bao, Y Wu, SH Liang, D Wang, LY Wu, et al., Tryptophan-kynure- nine metabolism: a link between the gut and brain for depression in inflam- matory bowel disease, J Neuroinflammation 18 (1) (2021) 135