Available online at www.sciencedirect.com

(2009) 420–425

Clinical Biochemistry 42

More rapid and sensitive method for simultaneous determination oftryptophan and kynurenic acid by HPLC

Pi Lan-gan a,b, Tang Ai-guo a,⁎, Mo Xi-ming a, Luo Xi-bo a, Ao Xiang a

a Department of Clinical Laboratory, The Second Xiangya Hospital, Central South University, No.139, Renmingzhong Road, Changsha, Hunan 410011, Chinab Department of Clinical Laboratory, The First People's Hospital, Chenzhou, Hunan 423000, China

Received 4 August 2008; received in revised form 14 October 2008; accepted 12 November 2008Available online 3 December 2008

Abstract

Objective: To describe a simple, rapid, and sensitive HPLC method for simultaneous determination of TRP and KYNA in human serum.Design and method: Samples were deproteinized by perchloric acid. KYNAwas detected at 344 nm of excitation wavelength and 404 nm of

emission wavelength, TRP was detected at 254 nm and 404 nm, with a total run time of 13 min per sample.Results: Standard curves of 0.49 μmol/L to 196 μmol/L of TRP were linear. Inter-day and intra-day coefficient of variations were 3.31% and

4.14%, respectively. Average recovery was 104.43%. Detection limit was 0.001 μmol/L. The linearity of the assay was maintained from 1.5 nmol/L to 2093 nmol/L of KYNA. Inter-day and intra-day CVS were 3.20% and 4.27%, respectively. Average recovery was 101.19%. Detection limitwas 0.05 nmol/L.

Conclusion: The developed HPLC method is simple, convenient and can be applied in the diagnosis of related diseases.© 2008 The Canadian Society of Clinical Chemists. Published by Elsevier Inc. All rights reserved.

Keywords: HPLC; On-column derivatization; Tryptophan; Kynurenic acid; Dual-wavelength; Fluorescence detection; Human serum

Introduction

L-Tryptophan (TRP) is an essential amino acid and anabundant protein component. In mammals, TRP is mostlymetabolized via the kynurenine (KYN) pathway in whichtryptophan is catabolised into kynurenine [1]. Then kynurenineis catabolised into kynurenic acid (KYNA) that acts as anendogenous antagonist of ionotropic glutamate receptors [2,3].KYNA can bind to the glycine-binding site of the N-methyl-D-aspartate (NMDA) receptor to antagonize its function [4]. It alsoshows antagonistic activity with respect to α7 nicotinicacetylcholine receptor [5]. Therefore, endogenous KYNAplays important roles in brain diseases [6–9], and is involvedin various neurological diseases [10–17]. Serum TRP levels arealso reported to undergo abnormal changes in many KYNA-associated diseases [10]. Therefore, it is very important tomeasure the concentrations of TRP and KYNA in vital fluids tomake a proper diagnosis. So far, only few methods reported the

⁎ Corresponding author. Fax: +86 0731 5292142.E-mail address: aiguotang@yahoo.com.cn (T. Ai-guo).

0009-9120/$ - see front matter © 2008 The Canadian Society of Clinical Chemistsdoi:10.1016/j.clinbiochem.2008.11.011

simultaneous determination of TRP and KYNA in serum withon-column derivatization [18]. Based on previous workincluding ours [18–21], here we report the development of amore sensitive and convenient HPLC method to simultaneouslydetermine TRP and KYNA in serum.

Materials and methods

Instrumentation and reagents

The HPLC system consisted of a Waters 510 pump, a Waters2475 fluorescence detector, a data processing system (HS2000,Hangzhou Yingpu Corp, Hangzhou, China) and a RheodyneModel 7725 injection valve with a 20 μL sample fixed loop(Rheodyne, Cotati, CA, USA). Other equipments contained aMillipore filter with 0.45 μm filter membrane (Millipore, USA),a Milli-Q purification system (Millipore, USA), TGLL-18 high-speed refrigerated desktop centrifuge (Taicang, China).

KYNA, TRP, kynurenine (KYN), phenylalanine (PHE),tyrosine (TYR), and creatinine (CRE) were purchased fromSigma (St. Louis, MO, USA). Acetonitrile and methanol were

. Published by Elsevier Inc. All rights reserved.

421P. Lan-gan et al. / Clinical Biochemistry 42 (2009) 420–425

HPLC grade (Tedia, USA). Except for perchloric acid that wasguaranteed grade, all other chemicals including zinc acetate,sodium acetate and acetic acid were analytical grade. Allsolutions were prepared with ultrapure water purified by theMilli-Q purification system.

Chromatographic conditions

The HPLC system is equipped with a Hypersil C8 column(300 mm×6.0 mm i.d. 10 μm particle size, Elite AnalyticalInstruments, Dalian), China. The mobile phase consisted of0.5 mol/L zinc acetate, 50 mmol/L sodium acetate with 6% (v/v)acetonitrile. At the beginning of the run, fluorescence excitationand emission wavelengths were operated at 344 nm and404 nm, respectively, and the excitation was changed to 254 nm9.5 min later. Flow rate was kept at 1.5 mL/min. All operationswere carried out at room temperature.

Reagent preparation

Perchloric acid in 5% (v/v) aqueous solution was used as theprotein precipitant. Stock solutions of PHE, TYR, 5-HT, andCRE were prepared in 2.5% (v/v) perchloric acid individuallyand were used as interference samples. The standard stocksolution of KYNA (104.67 μmol/L) was prepared with waterand TRP (4.900 mmol/L) was prepared in 2.5% (v/v) perchloricacid. Both solutions were stored at −30 °C. A standard mixtureof KYNA (final concentration, 26.17 nmol/L) and TRP (final24.5 μmol/L) TRP was prepared by mixing the two freshlythawed stock solutions in 2.5% (v/v) perchloric acid.

Sample collection and preparation

A total of 108 healthy subjects (55 males and 53 females;aged between 18–58 years old, y, mean age: 32.2±9.0 years)were recruited, the males aged between 18–54 years (meanage: 32.2±9.8 years) and the females, 21–58 years (mean age:32.1±8.1 years). Those having a history of past psychiatricillness, diagnosed autoimmune disease, or drug or alcoholabuse were excluded. All subjects were free from any physicalillness for at least 2 weeks before the study. They showednormal laboratory findings in blood chemistry (blood glucose,cholesterol, renal, thyroid and liver function), Venereal DiseaseResearch Laboratory tests, chest X-ray, and ECG. All subjectswere given written informed consent on entry into this study,and the protocol was approved by the Ethics Committee of theSecond Xiangya Hospital, Central South University.

After overnight fasting, 2 mLof venous blood was drawnfrom each participant. The collected samples were centrifugedat 2000 ×g for 10 min within 30 min of collection and the serumsamples were stored at −30 °C before analysis. Frozen serumwas thawed at room temperature, and the thawed sample wasdeproteinized by adding equal volume of 5 %(v/v) perchloricacid. The acidified serum was vortexed, stood at roomtemperature for 10 min to precipitate the protein, andcentrifuged for 10 min at 9000 ×g. Twenty microliters of thesupernatant was injected into HPLC column for analysis.

Qualitative and quantitative analysis

The existence of KYNA and TRP in serum was determinedby comparison and superposition of peak retention time toexternal standards. The concentrations of KYNA and TRP inserum were determined from their peak areas relative to anexternal standard according to the equation: KYNA or TRP(nmol/L or μmol/L)= (the peak area of KYNA or TRP inserum) / (the peak area of KYNA or TRP in standardsolution)× (the concentration of KYNA or TRP in standardsolution)×2. Data acquisition and processing were performedon the HS2000 data processing system.

Statistical analysis

Normally distributed data between groups were analyzedusing Student's t-test. When the data were not distributednormally, the analyses were performed using non-parametricMann-Whitney U test. Correlations were assessed by themethod of Spearman. Analyses were performed using SPSSversion 10.05 software package. Significance was set at a pvalue of less than 0.05.

Results

Selection of the excitation and emission wavelengths

Each of the standard solutions of TRP (24.5 μmol/L) andKYNA (26.17 nmol/L) were prepared separately and scanned ina range of 200–400 nm to determine the maximal excitationwavelength and in a range of 300–500 nm to determine themaximal emission spectrum, respectively, by using the Model2475 fluorescence detector with a manual stop-flow technique.Results showed that the peak wavelengths of excitation andemission of the TRP were at 254 nm and 404 nm, respectively,and KYNA were at 344 nm and 404 nm, respectively. Thesepeak wavelengths were used in the following experiments.

Selection of zinc acetate concentration in the mobile phase

To optimize the zinc acetate concentration used in the mobilephase for KYNA and TRP fluorescent derivatization, wemeasured the fluorescence intensities of KYNA and TRP inmobile phases with zinc acetate concentrations varying from 0 to0.5mol/L. As seen from Fig. 1, the peak areas of KYNA increasewith an increase of the concentration of zinc acetate. In contrast,the peak areas of TRP were barely changed as the concentrationof zinc acetate increases up to 0.5mol/L. To avoid small particlesof zinc acetate crystallized at higher concentrations due to its lowdissolving capability in the mobile phase, we chose the mobilephase containing 0.5 mol/L zinc acetate.

Effect of acetonitrile in the mobile phase on the retention timesof KYNA and TRP

To optimize the retention condition of resolving KYNA andTRP in their own standard working solution or as a mixture, the

Fig. 3. Effect of the flow rates on the retention time of KYNA (diamonds) andTRP (square). The mobile phase was pH 6.2 and contained 0.5 mol/L zincacetate, 50 mmol/L sodium acetate with 6% (v/v) acetonitrile. Data points weresimply connected by curves without any regression treatments.

Fig. 1. Dependence of the fluorescence intensity of KYNA and TRP on zincacetate concentrations. Fluorescence measurements on HPLC-FLD were madeat 404 nm with 254 nm excitation for TRP (diamonds) or at 404 nm with 350 nmexcitation for KYNA (triangles). The mobile phase contained 50 mmol/Lsodium acetate and 6% (v/v) acetonitrile; flow rate: 1.5 mL/min. Data pointswere simply connected by curves without any regression treatments.

422 P. Lan-gan et al. / Clinical Biochemistry 42 (2009) 420–425

concentrations of acetonitrile in the mobile phase were variedfrom 2.0% to 8.0%. As shown in Fig. 2, increasing theacetonitrile concentration in the mobile phase decreases theretention time of both KYNA and TRP. While 2.0% acetonitrileyielded better resolution of either TRP or KYNA in serum, itneeded a significantly long time. We found that both KYNA andTRP were resolved well at 6.0% acetonitrile within a reasonableretention time of 14 min. A mobile phase containing 7.0% and8.0% acetonitrile did not led to satisfactory separation of TRP inserum, and a mobile phase of 5.0% acetonitrile yielded poorseparation of KYNA in serum.

Effect of flow rates on retention times of KYNA and TRP

A series of flow rates from 0.5 to 2.0 mL/min were examinedto optimize the retention time of resolving KYNA and TRP intheir own standard working solution or as a mixture. The resultis shown in Fig. 3. The retention time of KYNA and TRP wasboth decreased as the flow rates increased. Meanwhile, the peakareas of KYNA and TRP were also decreased as the back-pressure of chromatographic system increased. While a flowrate of 0.5 mL/min or 1.0 mL/min gave a too long retentiontime. A flow rate of 2.0 mL/min led to unsatisfactory separationof TRP in serum. Hence we chose 1.5 mL/min as the flow ratesince each of KYNA and TRP could be eluted with sharp andsymmetric peaks within shorter retention times.

Fig. 2. Dependence of the retention time of KYNA (diamonds) and TRP(squares) on the acetonitrile concentration. The mobile phase was pH 6.2 andcontained 0.5 mol/L zinc acetate, 50 mmol/L sodium acetate; flow rate: 1.5 mL/min. Data points were simply connected by curves without any regressiontreatments.

Selection of pH of the mobile phase

The standard solution of KYNA or TRP was analyzed atdifferent pH values of the mobile phase. The fluorescenceintensities exhibited by KYNA and TRP were pH dependent,increasing with increasing pH (Table 1). The peaks of KYNA andTRP were thoroughly overlapped when the mobile phase pH wasat 4.5 and 5.0, partly overlapped at pH 5.5, and absolutelyseparated at pH 6.2.Moreover, the fluorescent complex of KYNAwith zinc ion was preferentially formed at pH 6.2. The pH of themobile phase containing 0.5 mol/L zinc acetate, 50 mmol/Lsodium acetate with 6% (v/v) acetonitrile was 6.2, which avoidsthe tedious work of adjusting the pH of mobile phase.

Chromatograms of the standards and human serum sample

A chromatographic analysis of an aqueous standard solutionis presented in Fig. 4A and a typical chromatographic analysisof human serum in Fig. 4B. The retention times of KYNA andTRP were respectively 8.1 and 11.3 min.

Linearity and detection limits

The linearity and detection limit were determined byanalyzing a series of standard solutions of KYNA from 0.01to 2618 nmol/L and of TRP from 0.0005 to 245 μmol/L. Theirsignals were proportional to their concentrations within theseranges. Correlation and regression were analyzed by the least-squares method, with correlation coefficients of 0.9997 forKYNA and 0.9999 for TRP. The detection limits that weredetermined under the conditions of the signal-to-noise ratio notless than 3 and an injection volume of 20 μL are presented in

Table 1Effects of the pH on retention and peak area

pH KYNA (52.34 nmol/L.) TRP (49 μmol/L.)

Retentiontime/min

Peakarea/μV×s

Retentiontime/min

Peakarea/μV×s

4.5 8.268 45278 7.838 29968815.0 9.282 122194 8.983 48026905.5 9.17 276600 10.078 64035306.2 8.075 534965 11.20 7727689

Table 2Regression analysis of calibration curves and detection limits of KYNA andTRP

Compound Regressionequation ⁎

r Linearrange

Detectionlimit ⁎⁎

KYNA (nmol/L.) Y=10472X+338.88 0.9996 1.05–2093 0.05TRP (μmol/L.) Y=5673+160526 X 0.9999 0.49–196 0.001

* X is concentration; Y is peak area, μV×s.** Detection limits were determined based on signal-to-noise ratio of 3(S/N=3).

423P. Lan-gan et al. / Clinical Biochemistry 42 (2009) 420–425

Table 2. Our detection limits, particularly for KYNA, werelower than those published data [11,18,19] and the linear rangeswere also extended as compared to those methods.

Precisions

As shown in Table 3, the precision of the method wasestimated by measuring the intra-day and inter-day reproduci-bility of serum specimens. The intra-day and inter-dayvariations were both less than 5% for KYNA and TRP.

Recoveries

Recovery was examined using serum specimens containing27.22 nmol/L KYNA and 53.50 μmol/L TRP. Each sample wastested in triplicate. The results are shown in Table 4.

Fig. 4. Chromatographic analysis of an aqueous standard solution (A) and a typicalcolumn (300 mm×6.0 mm i.d. 10 μm); mobile phase: 0.5 mol/L zinc acetate, 50 mmwavelength: 0–9.5 min 344 nm, 9.5 min later 254 nm; emission wavelength: 404 nmand TRP, respectively.

Interference test

To verify whether some substances in serum interfere withthe separation of KYNA and TRP, individual standard solutionsof PHE, TYE, KYNA and CRE were prepared. Each of them

chromatographic analysis of human serum (B). Other conditions: Hypersil C8ol/L sodium acetate with 6% (v/v) acetonitrile; flow rate: 1.5 mL/min; excitation; injection volume: 20 μl. 8.1 min and 11.3 min are the retention time of KYNA

Table 3Precisions of intraday and interday determination of KYNA and TRP

Component Intra-day (n=20) Inter-day (n=20)

(mean±SD) RSD/% (mean±SD) RSD/%

KYNA (nmol/L) 27.22±0.87 3.20 27.37±1.17 4.27TRP (μmol/L) 53.50±1.77 3.31 53.36±2.21 4.14

Table 5Result of interference test

Component Concentration Rentntion time/min PeakArea/μV×Sec

CRE (μmol/L) 442.0 – –TYR (μmol/L) 550.0 – –PHE (μmol/L) 610.0 – –KYN (μmol/L) 19.6 – –KYNA (nmol/L) 26.17 8.223 254151TRP (μmol/L) 24.5 11.29 3413871

Table 6Concentrations of serum KYNA and TRP in males and females (xFs)

Males (n=55) Females (n=53) P

KYNA (nmol/L) * 23.94±7.26 22.63±9.31 0.110TRP (μmol/L) 51.28±8.19 43.76±11.17 b0.001KA/T ratio (nmol/μmol) * 0.47±0.14 0.54±0.24 0.368

⁎ Statistical comparisons were performed using Mann-Whitney test.

424 P. Lan-gan et al. / Clinical Biochemistry 42 (2009) 420–425

was mixed with KYNA and TRP to do HPLC analysis. Theresults of retention time and peak area were summarized inTable 5. Based on the results, we concluded that none of theminterfered with the method developed above.

KYNA and TRP in serum of Chinese healthy person

The serum KYNA and TRP values among 108 Chinesehealthy persons were 11.85–45.32 nmol/L. and 27.15–68.03 μmol/L, respectively, and the value of KA/T ratio(nmol/μmol) was 0.22∼1.03 (Table 6). Among the 108 healthypersons, TRP is correlated with KYNA and age (rs=0.192,P=0.046 and rs= -0.247, P=0.010). The KA/T ratio is alsocorrelated positively with age (rs=0.333, Pb0.001). No genderdifference was found in the concentrations of KYNA or the KA/T ratios (PN0.05). Moreover, when health subjects weredivided into two groups with ages≤30 years and N30 years,the latter group showed significantly lower TRP concentrationsand higher KA/T ratios. However, such age difference was notseen in serum KYNA levels (Table 7). No correlation was foundbetween serum KYNA levels and the age of the subject.

Discussion

Several chromatographic protocols have been presented todetermine TRP and KYNA, including the application of HPLCdetected by fluorescence [20,21], electrochemical signals [22],mass spectrometry [23], or a combination of these threetechniques [18,19]. The sensitivity of HPLC with UV is toolow for biological sample. Most of these methods requiregradient elution or multi-step detection and are time consuming.Considering the requirement of 0.5 mLserum by one method[18], it is preferable to develop a more sensitive method using asmaller amount of serum to simultaneously determine TRP andKYNA. Previously, the method with fluoresce detection ofKYNA used a mobile phase of pH 4.9 instead of 6.2, though thelatter was preferred to form a fluorescent complex of KYNAwith zinc ion [19]. The reason for not using pH 6.2 was that aconsiderably small capacity factor of KYNA on an octadecyl

Table 4Recoveries of KYNA and TRP

Component Added Found Recovery/% Average recovery/%

KYNA (nmol/L) 10.47 10.89 104.01 101.1920.93 21.68 103.5852.33 50.22 95.97

TRP (μmol/L) 9.8 10.65 108.88 104.4324.5 25.90 105.7149.0 46.90 98.71

silica (ODS) column at this pH led to insufficient separation ofKYNA [19]. In order to solve these problems, Mitsuhashi [11]developed a column-switching HPLC system consisting of twocolumns in series to separate KYNA more efficiently. However,this method was costly and unsuitable for routine clinical use.

Zinc acetate has been reported to greatly enhance thefluorescence of KYNA and used for the fluorescence detectionof KYNA [18]. In our method, the mobile phase also contained0.5 mol/L zinc acetate. We found that the fluorescence intensityof KYNA increased by increasing the concentration of zincacetate. When the mobile phase contained 0.5 mol/L zincacetate, the fluorescence intensity of KYNA was fifty-foldhigher than that without zinc acetate. By optimizing the zincacetate concentration to maximize the fluorescence intensity ofKYNA, we have greatly improved the limits of fluorescencedetection as compared with previous method [11,18]. Under theoptimized chromatographic conditions, the retention time ofTRP and KYNA was 8.1 min and 11.3 min, respectively,meaning that approximately 50 samples can be measured withone HPLC system per day. Compared to the previous methods[11,18], our method is faster without loss specificity andsensitivity. The presented method has the potential to be usedfor high throughput detection of samples.

In conclusion, we have developed a method for simultaneousdetermination of KYNA and TRP in human serum by HPLC-FLD with on-column fluorescence derivatization. The methodis simple and rapid, and its precision, sensitivity and accuracyare satisfactory. The method meets the requirements for routineanalysis of KYNA and TRP in serum. It will be also applicable

Table 7Concentrations of serum KYNA and TRP between age groups (xFs)

≤30 (n=53) N30 (n=55) P

KYNA (nmol/L) * 22.39±8.07 24.18±8.53 0.261TRP (μmol/L) 50.19±10.14 45.08±10.17 0.010KA/T ratio (nmol/μmol) * 0.46±0.18 0.55±0.20 0.006

⁎ Statistical comparisons were performed using Mann-Whitney test.

425P. Lan-gan et al. / Clinical Biochemistry 42 (2009) 420–425

to determine KYNA and TRP in brain tissue or microdialysatesample for the clinical and basic researches.

Acknowledgments

This research was supported by Science and TechnologicalDepartment of Hunan Province (No. 05SK3026) and publicHealth Department of Hunan Province (No.C2005019). Wethank Dr. Guoqing Tang for his continuous support.

References

[1] Stone TW, Darlington LG. Endogenous kynurenines as targets for drugdiscovery and development. Nat Rev Drug Discov 2002;1:609–20.

[2] Swartz KJ, During MJ, Freese A, Beal MF. Cerebral synthesis and releaseof kynurenic acid: an endogenous antagonist of excitatory amino acidreceptors. J Neurosci 1990;10:2965–73.

[3] Stone TW. Inhibitors of kynurenine pathway. Eur J Med Chem 2000;35:179–86.

[4] Hilmas C, Pereira EF, Alkondon M, Rassoulpour A, Schwarcz R,Albuquerque EX. The brain metabolite kynurenic acid inhibitsα7 nicotinicreceptor activity and increases non-α7 nicotinic receptor expression:physiopathological implications. J Neurosci 2001;21:7463–73.

[5] Alkondon M, Pereira EF, Yu P, Arruda EZ, Almeida LE, Guidetti P, et al.Targeted deletion of the kynurenine aminotransferase II gene reveals acritical role of endogenous kynurenic acid in the regulation of synaptictransmission via alpha7 nicotinic receptors in the hippocampus. J Neurosci2004;24:4635–48.

[6] Erhardt S, Engberg G. Increased phasic activity of dopaminergic neuronesin the rat ventral tegmental area following pharmacologically elevatedlevels of endogenous kynurenic acid. Acta Physiol Scand 2002;175:45–53.

[7] Erhardt S, Schwieler L, Emanuelsson C, Geyer M. Endogenous KynurenicAcid Disrupts Prepulse Inhibition. Biol Psychiatry 2004;56:255–60.

[8] Carpenedo R., Pittaluga A., Cozzi A., Attucci S., Galli A., Raiteri M., etal., Presynaptic kynurenate-sensitive receptors inhibit glutamate release.Eur. J. Neurosci 2001;13:2141–7.

[9] Wu HQ, Fuxe K, Schwarcz R. Neuronalα1 receptors mediate increase inextracellular kynurenic acid after local intrastriatal adenosine infusion. JNeurochem 2004;90:621–8.

[10] Ravikumar A, Deepadevi KV, Arun P, Manojkumar V, Kurup PA.Tryptophan and tyrosine catabolic pattern in neuropsychiatric disorders.Neurol India 2000;48:231–8.

[11] Mitsuhashi S, Fukushima T, Kawai J. Improved method for thedetermination of kynurenic acid in rat plasma by column-switchingHPLC with post-column fluorescence detection. Anal Chim Acta2006;562:36–43.

[12] Erhardt S, schwieler L, Emanuelsson C, Geyer M. Endogenous KynurenicAcid Disrupts Prepulse Inhibition. Biol Psychiatry 2004;56:255–60.

[13] Nilsson LK, Linderholm KR, Engberg G, Paulson L, Blennow K,Lindström LH, et al. Elevated levels of kynurenic acid in thecerebrospinal fluid of male patients with schizophrenia. Schizophr Res2005;80:315–22.

[14] Myint AM, Kim YK, Verkerk R, Scharpé S, Steinbusch H, Leonard B.Kynurenine pathway in major depression: Evidence of impairedneuroprotection. J Affect Disord 2007;98:143–51.

[15] Hartai Z, Juhasz A, Rimanoczy A, Janáky T, Donkó T, Dux L, et al.Decreased serum and red blood cell kynurenic acid levels in Alzheimer'sdisease. Neurochem Int 2007;50:308–13.

[16] Hartai Z, Klivenyi P, Janaky T, Penke B, Dux L, Vecsei L. Kynureninemetabolism in plasma and in red blood cells in Parkinson's disease. JNeurol Sci 2005;239:31–5.

[17] Sapko MT, Guidetti P, Yu P, Tagle DA, Pellicciari R, Schwarcz R.Endogenous kynurenate controls the vulnerability of striatal neurons toquinolinate: Implications for Huntington's disease. Exp Neurol 2006;1971:31–40.

[18] Herve C, Beyne P, Jamault H, Delacoux E. Determination of tryptophanand its kynurenine pathway metabolites in human serum by high-performance liquid chromatography with simultaneous ultraviolet andfluorimetric detection. J Chromatogr, B, Biomed Appl 1996;675:157–61.

[19] Swartz KJ, Matson WR, MacGarvey U, Ryan EA, Beal MF. Measurmentof kynurenic acid in mammalian brain extracts and cerebrospinal fluid byhigh-performance liquid chromatography with fluorometric and coulo-metric electrode array detection. Anal Biochem 1990;185:363–76.

[20] Luo XB, Tang AG, Pi LG, Xiao LD, Ao X, Pu YH, et al. Determination ofkynurenine in serum by high performance liquid chromatography with on-column fluorescence derivatization. Clin Chim Acta 2008;389:186–8.

[21] Xiao LD, Luo XB, Pi LG, Tang AG. Simultaneous determination ofkynurenine and kynurenic acid concentrations in human serum by HPLCwith dual wavelengths fluorescence detection. Clinica Chimica Acta2008;395:178–80.

[22] Beal MF, Matson WR, Swartz KJ, Gamache PH, Bird ED. Kynureninepathway measurements in Huntington's disease striatum: evidence forreduced formation of kynurenic acid. J Neurochem 1990;55:1327–39.

[23] Morita I, Kawamoto M, Hattori M, Eguchi K, Sekiba K, Yoshida H.Determination of tryptophan and its metabolites in human plasma andserum by high-performance liquid chromatography with automated sampleclean-up system. J Chromatogr 1990;6;526:367–74.