- Email: [email protected]
Measurement of pyridoxal 5′-phosphate, pyridoxal, and 4-pyridoxic acid in the cerebrospinal fluid of children
Clinica Chimica Acta 466 (2017) 1–5
Contents lists available at ScienceDirect
Clinica Chimica Acta journal homepage: www.elsevier.com/locate/clinchim
Measurement of pyridoxal 5′-phosphate, pyridoxal, and 4-pyridoxic acid in the cerebrospinal fluid of children Tomoyuki Akiyama a,⁎, Mari Akiyama a, Yumiko Hayashi a, Takashi Shibata a,b, Yoshiyuki Hanaoka a,b, Soichiro Toda c, Katsumi Imai d, Shin-ichiro Hamano e, Tohru Okanishi f, Harumi Yoshinaga a,b, Katsuhiro Kobayashi a,b a
Department of Child Neurology, Okayama University Hospital, Okayama, Japan Department of Child Neurology, Okayama University Graduate School of Medicine, Dentistry and Pharmaceutical Sciences, Okayama, Japan c Department of Pediatrics, Kameda Medical Center, Chiba, Japan d Department of Pediatrics, NHO Shizuoka Institute of Epilepsy and Neurological Disorders, Shizuoka, Japan e Division of Neurology, Saitama Children's Medical Center, Saitama, Japan f Department of Child Neurology, Seirei Hamamatsu General Hospital, Hamamatsu, Japan b
a r t i c l e
i n f o
Article history: Received 6 November 2016 Received in revised form 24 December 2016 Accepted 27 December 2016 Available online 28 December 2016 Keywords: Pyridoxal phosphate 4-pyridoxic acid Vitamin B6 Liquid chromatography Fluorescence detection Epilepsy
a b s t r a c t Background: We quantified pyridoxal 5′-phosphate (PLP), pyridoxal (PL), and 4-pyridoxic acid (PA) in the cerebrospinal fluid (CSF) of children and to investigate the effect of age, sex, epilepsy, and anti-epileptic drug (AED) therapy on these vitamers. Methods: CSF samples prospectively collected from 116 pediatric patients were analyzed. PLP, PL, and PA were measured using high-performance liquid chromatography with fluorescence detection, using pre-column derivatization by semicarbazide. Effects of age, sex, epilepsy, and AEDs on these vitamers and the PLP/PL ratio were evaluated using multiple linear regression models. Results: The PLP, PL, and PA concentrations were correlated negatively with age and the PLP/PL ratio was correlated positively with age. Multiple regression analysis revealed that the presence of epilepsy was associated with lower PLP concentrations and PLP/PL ratios but sex and AED therapy had no influence on these values. The observed ranges of these vitamers in epileptic and non-epileptic patients were demonstrated. Conclusions: We showed the age dependence of PLP and PL in CSF from pediatric patients. Epileptic patients had lower PLP concentrations and PLP/PL ratios than non-epileptic patients, but it is unknown whether this is the cause, or a result, of epilepsy. © 2016 Elsevier B.V. All rights reserved.
1. Introduction Vitamin B6 consists of various vitamers, including pyridoxal (PL), pyridoxamine, pyridoxine, their phosphorylated forms (pyridoxal 5′phosphate [PLP], pyridoxamine 5′-phosphate, and pyridoxine 5′-phosphate), and 4-pyridoxic acid (PA). PLP, a biologically active form of vitamin B6, is involved in numerous biochemical reactions as a cofactor [1]. PLP is an essential cofactor of glutamate decarboxylase, which converts glutamic acid, an excitatory neurotransmitter, to γ-aminobutyric acid, a main inhibitory neurotransmitter in the central nervous system. Lack of PLP results in intractable epileptic seizures, such as neonatal epileptic encephalopathy caused by pyridox[am]ine-phosphate oxidase (PNPO)
Abbreviations: AED, anti-epileptic drugs; PA, 4-pyridoxic acid; PDXK, pyridoxal kinase; PL, pyridoxal; PLP, pyridoxal 5′-phosphate; PNPO, pyridox[am]ine-phosphate oxidase. ⁎ Corresponding author at: Department of Child Neurology, Okayama University Hospital, 2-5-1 Shikata-cho, Kita-ku, Okayama, 700-8558, Japan. E-mail address: [email protected] (T. Akiyama).
http://dx.doi.org/10.1016/j.cca.2016.12.027 0009-8981/© 2016 Elsevier B.V. All rights reserved.
deficiency [2]. Insufficient PLP also affects monoamine neurotransmitter metabolism because aromatic L-amino acid decarboxylase, an enzyme to synthesize dopamine and serotonin from their precursors, is dependent on PLP. There are few reports on vitamin B6 in cerebrospinal fluid (CSF), especially for children. Reference values of PLP in children were reported using high-performance liquid chromatography (HPLC) with fluorescence detection [3,4]. Other vitamers have been measured using liquid chromatography-tandem mass spectrometry, which showed that PLP and PL are the main forms of vitamin B6 in the CSF [5,6]. Reference values for CSF PL and PA in children 1 year of age or older were reported in only one study [6]. The impact of some factors, such as sex, epilepsy, and anti-epileptic drugs (AEDs), on these vitamer concentrations needs to be taken into account to interpret the assay results correctly. The negative effect of AEDs on the CSF PLP and PL concentration, and lower PL concentrations in males have been reported in children [6]. However, another study reported no effect of AEDs or seizures on CSF PLP concentrations [4].
2
T. Akiyama et al. / Clinica Chimica Acta 466 (2017) 1–5
In this study, we quantified the PLP, PL, and PA concentrations in CSF obtained from pediatric patients to investigate the effect of age, sex, epilepsy, and AEDs on the concentrations of these vitamers and the PLP/PL ratio. We also compared these values with reference values in the literature. 2. Patients and methods 2.1. Patients CSF samples were prospectively collected between October 2014 and November 2016 from pediatric patients (1 month to 17 y) who underwent lumbar puncture to investigate neurological symptoms (e.g. epileptic seizures, developmental delay) or suspected meningitis. We excluded patients with: 1) vitamin B6 supplementation (highdose therapy or supplement intake other than regular food); 2) CSF samples not protected from light within 15 min; 3) metabolic diseases that may affect vitamin B6 metabolism; 4) acute neurological process such as acute encephalitis/encephalopathy, febrile seizures, and status epilepticus; 5) significant blood contamination; and 6) statistical outliers. This study was approved by the ethics committee of Okayama University Hospital. Written informed consent was obtained from the parents or guardians of all the patients before the procedure.
The HPLC system consisted of a Waters Alliance 2695 module with a Waters 2475 multi λ fluorescence detector (Waters). The mobile phase A was 60 mmol/l sodium phosphate buffer with EDTA-2Na at an approximate pH of 6.5 (18 mmol/l disodium hydrogen phosphate, 42 mmol/l sodium dihydrogen phosphate, and 0.4 g/l EDTA-2Na), the mobile phase B was 100% acetonitrile, and the mobile phase C was ultrapure water. A gradient mode was used (Table 1), and the injection volume was 30 μl. The separation was performed at a flow rate of 1.2 ml/min through a reverse-phase column (Atlantis T3, 3 μm, 3.0 × 50 mm, Waters Japan) at 35 °C with a guard column (Atlantis T3 VanGuard Cartridge, 3 μm, 3.9 × 5 mm, Waters). The PLP and PL derivatives were detected using excitation (Ex) at 370 nm and emission (Em) at 460 nm. PA was detected by its natural fluorescence (Ex 320 nm, Em 420 nm). Concentrations were calculated using area under the curve, with a single-point calibrator at 1024 nmol/l. The entire analysis time was 20 min.
2.4. Statistical analysis Statistical analysis was performed using JMP 4.0 (SAS Institute). Correlation was evaluated using Spearman's correlation coefficient. The independent effects of age, sex, epilepsy, and AEDs were tested using multiple regression analysis. The significance level was set to 0.05.
2.2. CSF collection protocol Because no rostrocaudal gradient of B6 vitamers has been reported [5], we used any fraction of the CSF samples in this study. Collected CSF samples were protected from light within 15 min and frozen within 2 h. The stability of CSF B6 vitamers for 10 h and plasma PLP for 12 h at room temperature was reported when the samples were kept in the dark [5,7]. They were stored at −80 °C until analysis. When the samples were collected outside Okayama University Hospital, they were shipped on dry ice to our laboratory. 2.3. Measurement of PLP, PL, and PA We obtained PLP monohydrate (82870), PL hydrochloride (P9130), PA (P9630), and ethylenediaminetetraacetic acid disodium salt (EDTA-2Na) dihydrate (09-1420) from Sigma Aldrich (St. Louis, MO, USA), sodium dihydrogen phosphate (197-09705), disodium hydrogen phosphate (197-02865), semicarbazide hydrochloride (192-00372), and glycine (073-00732) from Wako Pure Chemical Industries, Ltd. (Osaka, Japan). Acetonitrile (1.00030.2500) was from Merck Millipore (Billerica, MA, USA). Ultrapure water was prepared using a Direct-Q UV3 system (Merck Millipore). Simultaneous measurement of PLP, PL, and PA was conducted according to Talwar et al. [7] with some modifications. Primary stock solutions of PLP, PL, and PA were prepared, mixed, and diluted with ultrapure water to achieve final concentrations of 250 μmol/l for each component. This secondary stock solution was further diluted with ultrapure water to 1024 nmol/l to make a single-point calibrator. Quality control samples at 64 nmol/l were also prepared. Derivatization of samples was conducted as described below. Briefly, 200 μl of undiluted CSF samples or 100 μl of serum/plasma samples diluted two-fold with 100 μl ultrapure water were derivatized by adding 16 μl of derivatization agent (370 mg/ml of semicarbazide hydrochloride and 250 mg/ml of glycine dissolved in ultrapure water). After vortexing and incubating the samples at 40 °C in the dark for 30 min, 16 μl of 70% perchloric acid was added and vortexed for deproteinization. After centrifuging for 5 min at 16,000 × g, 170 μl of supernatant was transferred to a new microtube and 23 μl of 25% sodium hydroxide was added to adjust the pH to 3.0–5.0. The samples were then filtered through 0.45-μm Millex LH filters (Merck Millipore) and subjected to HPLC.
3. Results 3.1. Patient characteristics There were 178 patients who underwent lumbar puncture to measure PLP, PL, and PA concentrations in the CSF. Sixty-two patients were excluded for the following reasons: 1) 40 patients with CSF samples protected from light N 15 min after collection; 2) 9 patients with acute neurological process; 3) 8 patients with metabolic diseases; 4) 2 patients on high-dose PLP therapy; and 5) 3 patients considered to be statistical outliers, whose PLP, PL, and PA concentrations were all above the cut-off values (3rd quartile + 1.5 × interquartile range). The data from the remaining 116 patients (68 males and 48 females) were used for analysis in this study. Patient age ranged from 1 to 199 months (median age, 48 months). There were 62 patients with a diagnosis of epilepsy (median age, 32.5 months), 50 of whom were taking AEDs. There were 54 patients without epilepsy (median age, 75.5 months), 6 of whom were taking AEDs. For these 6 patients, AEDs were administered for suspected epileptic seizures, or were used as muscle relaxants for dystonia or spasticity, or as mood stabilizers. In total, 56 patients were taking AEDs (median age, 45.5 months) and 60 were not taking AEDs (median age, 49.5 months). The administered AEDs were valproic acid in 29 patients, levetiracetam in 14 patients, zonisamide in 13 patients, phenobarbital in 9 patients, carbamazepine in 8 patients, topiramate in 8 patients, clonazepam in 8 patients, lamotrigine in 7 patients, clobazam in 7 patients, ethosuximide in 3 patients, potassium bromide in 2 patients, nitrazepam in 1 patient, and rufinamide in 1 patient either as monotherapy or in combination. Table 1 High-performance liquid chromatography gradient settings. Time (min)
Flow (ml/min)
Mobile phase A (%)
Mobile phase B (%)
Mobile phase C (%)
Gradient curve
Initial 3.5 7.0 7.1 10.0 20.0
1.2 1.2 1.2 1.2 1.2 1.2
98.6 98.6 94.0 0.0 98.6 98.6
1.4 1.4 6.0 60.0 1.4 1.4
0.0 0.0 0.0 40.0 0.0 0.0
Linear Linear Step Step Linear
T. Akiyama et al. / Clinica Chimica Acta 466 (2017) 1–5
3.2. Validity of PLP, PL, and PA assay using HPLC A representative chromatogram of a standard sample and a CSF sample is shown in Fig. 1. PLP, PA, and PL were eluted at 2.3 min, 3.7 min, and 7.2 min, respectively. The limit of detection (signal-to-noise [S/N] ratio ≥ 3) was 1.5 nmol/l for PLP, 1.0 nmol/l for PL, and 0.4 nmol/l for PA. The limit of quantification (LOQ, S/N ratio ≥ 10) was 3.5 nmol/l for PLP, 2.0 nmol/l for PL, and 1.2 nmol/l for PA. There was a good linear relationship over the concentration range of LOQ to 1024 nmol/l for PLP (R2 = 0.9996), PL (R2 = 0.9999), and PA (R2 N 0.9999). Because the regression line intercepts for these compounds were not significantly different from zero, we chose to use a single-point calibrator at 1024 nmol/l. The intraday coefficients of variation (CV) were 0.9% for PLP, 0.3% for PL, and 0.2% for PA (n = 5) for a quality control sample at 64 nmol/l. Inter-day CVs were 1.1% for PLP, 1.4% for PL, and 0.9% for PA (n = 5). PLP, PL, and PA recovery from the CSF samples spiked with 25.6 nmol/l (n = 3) was 95.7– 98.8% (mean, 96.8%), 100.1–101.4% (mean, 100.8%), and 96.6–98.0% (mean, 97.1%), respectively. PLP, PL, and PA recovery from the CSF samples spiked with 102.4 nmol/l (n = 3) was 98.5–100.2% (mean, 99.2%), 100.0– 100.1% (mean, 100.1%), and 99.3–101.7% (mean, 100.2%), respectively.
3
incorporating age, sex, presence of epilepsy, and AED therapy showed that presence of epilepsy (p = 0.0015) and age (p b 0.0001) were associated with lower PLP concentrations (Table 2). Sex and AED therapy did not affect the PLP concentration. The PL concentration was negatively correlated with age (ρ = −0.728, p b 0.0001; Fig. 2B). Multiple regression analysis showed that only age (p b 0.0001) influenced the PL concentration (Table 2). Sex, presence of epilepsy, and AED therapy did not affect the PL concentration. The PLP/PL ratio was positively correlated with age (ρ = 0.524, p b 0.0001; Fig. 2C). Multiple regression analysis demonstrated that age (p b 0.0001) and presence of epilepsy (p = 0.0427) independently influenced the PLP/PL ratio (Table 2). Sex and AED therapy did not affect the PLP/PL ratio. We did not perform a statistical analysis on PA, because most patients except for young patients had PA concentrations below the LOQ (b1.2 nmol/l; Fig. 2D). The observed ranges of PLP, PL, PA, and PLP/PL according to age groups and presence/absence of epilepsy are presented in the Table 3. 4. Discussion
3.3. PLP, PL, and PA concentrations, and the PLP/PL ratio The PLP concentration was negatively correlated with age (ρ = − 0.310, p = 0.0007; Fig. 2A). Multiple regression analysis
We quantified the PLP, PL, and PA concentrations, and the PLP/PL ratio in pediatric CSF samples and showed their dependence on age. Although sex and AED therapy did not affect the concentrations of these
Fig. 1. Chromatograms of a standard sample and a CSF sample. A) A standard sample of PLP, PL, and PA at 1024 nmol/l. B) A CSF sample from a 7-month-old patient. PLP, PL, and PA were well separated from other peaks (PLP, 28.1 nmol/l; PL, 52.9 nmol/l; and PA, 1.5 nmol/l). PLP, pyridoxal 5′-phosphate; PL, pyridoxal; PA, 4-pyridoxic acid; Ex., excitation; Em., emission.
4
T. Akiyama et al. / Clinica Chimica Acta 466 (2017) 1–5
Fig. 2. PLP, PL, and PA concentration and PLP/PL ratio in the cerebrospinal fluid according to age. A) PLP vs. age. PLP was correlated negatively with age. B) PL vs. age. PL was correlated negatively with age. C) PLP/PL ratio vs. age. PLP/PL ratio was positively correlated with age. D) PA vs. age. PA showed high concentrations at young ages. Filled circles, non-epileptic patients; open squares, epileptic patients; PLP, pyridoxal 5′-phosphate; PL, pyridoxal; PA, 4-pyridoxic acid.
vitamers, epileptic patients had lower PLP values and PLP/PL ratios than non-epileptic patients. The dependence of PLP on age (higher PLP values at a younger age) is consistent with past studies [3,4]. Our study showed a similar age
Table 2 Results of multiple regression analysis. Vitamers
Variables
p
PLP
Age Sex Epilepsy AED Age Sex Epilepsy AED Age Sex Epilepsy AED
b0.0001⁎ NS 0.0015⁎⁎
PL
PLP/PL
NS b0.0001⁎ NS NS NS b0.0001⁎⁎⁎ NS 0.0427⁎⁎ NS
PLP, pyridoxal 5′-phosphate; PL, pyridoxal; AED, anti-epileptic drugs; NS, not significant. ⁎ Negative correlation with age. ⁎⁎ Presence of epilepsy was associated with lower values. ⁎⁎⁎ Positive correlation with age.
dependence for PL concentration, which was not investigated in a previous study that measured CSF PL in children [6]. The degree of decline with age seems to be different between PLP and PL based on the age dependence of the PLP/PL ratio. The observed PLP, PL, and PA ranges and the PLP/PL ratio in non-epileptic patients were similar to previously reported values [3,4,6]. There is not as much information on the PL assay as on the PLP assay, and measurement of both PL and PLP may provide useful clinical information. Production of PLP in the central nervous system requires an appropriate supply of non-phosphorylated vitamin B6 vitamers, including PL, from the peripheral blood and sufficient activity of pyridoxal kinase (PDXK) and PNPO [1]. Because PLP is not membrane-permeable, PLP in the blood has to be dephosphorylated to PL by a membrane-bound tissue non-specific alkaline phosphatase [1]. After PL crosses the cell membrane from the blood, it is phosphorylated back to PLP by PDXK and trapped within the cell. Although the brain itself does not release PLP, the choroid plexus releases PLP into the CSF [8]. When CSF PLP is low, the PL assay may clarify whether it is caused by deficient phosphorylation. Some toxins (e.g. ginkgotoxin) and medications (e.g., isoniazid) are known to inhibit PDXK [9]. Association between CSF PLP values and AED therapy or epilepsy has not been consistently shown. Footitt et al. demonstrated no statistical difference of CSF PLP concentrations between patients with and without
T. Akiyama et al. / Clinica Chimica Acta 466 (2017) 1–5 Table 3 Observed ranges of PLP, PL, PA, and PLP/PL according to age. PLP (nmol/l) Patients without epilepsy 1–11 months 16.9–51.4 (n = 12) (median 37.2) 1–6 years (n = 15) 15.3–49.3 (24.5) 7–17 years (n = 27) 13.1–41.2 (22.6) Patients with epilepsy 1–11 months (n = 22) 1–6 years (n = 29) 7–17 years (n = 11)
5.9–56.3 (30.9) 8.7–50.8 (18.9) 6.6–32.0 (17.6)
PL (nmol/l)
PA (nmol/l)
PLP/PL
17.5–97.5 (51.6) 11.7–34.8 (26.4) 12.0–38.9 (17.8)
b1.2a–3.3 (1.4) b1.2a–1.3 (b1.2a) b1.2a
0.24–0.99 (0.53) 0.76–1.77 (1.08) 0.58–2.25 (1.10)
21.2–99.8 (47.4) 7.1–44.4 (25.0) 7.9–25.0 (16.3)
a
b1.2 –3.2 (1.3) b1.2a–2.1 (b1.2a) b1.2a
0.14–1.47 (0.56) 0.39–1.54 (0.75) 0.83–1.96 (1.01)
PLP, pyridoxal 5′-phosphate; PL, pyridoxal; PA, 4-pyridoxic acid. a Below the limit of quantification (1.2 nmol/l).
seizures, and between those with and without AED therapy [4]. Albersen et al. showed that PLP and PL concentrations were lower in patients with AEDs than in those not receiving AEDs but these values were not influenced by epilepsy [6]. The current study showed that there is a lower PLP concentration and PLP/PL ratio in epileptic patients. This effect was significant even when AED use is taken into account. A negative impact of epilepsy on the PLP concentration and PLP/PL ratio suggests that decreased PLP production may play a role in the pathophysiology of epilepsy or that low PLP values are the result of epilepsy, but it is difficult to distinguish between these two potential mechanisms at this point. Although the use of AEDs was not associated with PLP and PL in the current study, we did not have enough data to investigate the effect of each kind of AED. The majority of epileptic patients were treated with AEDs at the time of lumbar puncture, which makes it difficult to separate the effect of AED from that of epilepsy. We cannot fully exclude the possibility that certain AEDs that are used mainly for epileptic patients in this study had an effect on PLP and PL concentrations in the CSF. For PLP and PL measurement, we used pre-column derivatization by semicarbazide hydrochloride, which forms stable and strongly fluorescent derivatives, thereby making sample handling easy and enhancing the assay sensitivity. This derivatization process was reported to be complete within 30 min at the ambient temperature [7]. We chose to perform this reaction at 40 °C because we experienced lower and variable recoveries (80–95%) of PLP and PL with undiluted serum/plasma samples derivatized at the ambient temperature. This study has some limitations. A larger number of samples will clarify whether epilepsy or certain kinds of AEDs influence the CSF PLP concentration. Although we excluded patients with possible alterations in vitamin B6 metabolism, some patients may have unknown disorders that can affect the assay results. Because CSF PL is influenced by plasma PL [6], further study with simultaneous measurement of B6 vitamers in the CSF and plasma is necessary to understand the overall dynamics of these vitamers. We are currently in the process of collecting paired CSF and plasma samples for this purpose. 5. Conclusion We quantified the PLP, PL, and PA concentration in the CSF of pediatric patients with neurological symptoms. The PLP, PL, and PA
5
concentrations decreased with age but the PLP/PL ratio increased with the age. Patients with epilepsy had lower CSF PLP concentrations and PLP/PL ratios than those without epilepsy, which might be either the cause or a result of epilepsy. Use of AEDs did not affect these values. Observed ranges of PLP, PL, and PA concentrations were also demonstrated. Financial disclosure This study was supported by the Japan Society for the Promotion of Science (Grant Number JP15K09622). This funding source had no involvement in the study design, the collection, analysis, and interpretation of data, the writing of the report, or the decision to submit this article. Acknowledgements T.A. received grant support from the Japan Society for the Promotion of Science. We thank Dr. Shinpei Baba at Tokyo Medical and Dental University, Dr. Tetsuhiro Fukuyama at Nagano Children's Hospital, Dr. Tomohide Goto at Kanagawa Children's Medical Center, Dr. Yoshikazu Kitami at Tokyo Metropolitan Children's Medical Center, Dr. Masaya Kubota at National Center for Child Health and Development, Ichiro Kuki at Osaka City General Hospital, Dr. Satoko Kumada at Tokyo Metropolitan Neurological Hospital, Dr. Shinji Saito at Nagoya City University, Dr. Takashi Shiihara at Gunma Children's Medical Center, Dr. Kentaro Shirai at Tsuchiura Kyodo Hospital, Dr. Akihito Takahashi at Kurashiki Central Hospital, and Dr. Yoshihiro Toda at Tokushima University, for providing CSF samples from pediatric patients. We thank Eibunkousei. net (http://www.eibunkousei.net/) for English-language editing. References [1] R. Surtees, P. Mills, P. Clayton, Inborn errors affecting vitamin B6 metabolism, Future Neurol. 1 (2006) 615–620. [2] P.B. Mills, S.S. Camuzeaux, E.J. Footitt, K.A. Mills, P. Gissen, L. Fisher, K.B. Das, S.M. Varadkar, S. Zuberi, R. McWilliam, T. Stodberg, B. Plecko, M.R. Baumgartner, O. Maier, S. Calvert, K. Riney, N.I. Wolf, J.H. Livingston, P. Bala, C.F. Morel, F. Feillet, F. Raimondi, E. Del Giudice, W.K. Chong, M. Pitt, P.T. Clayton, Epilepsy due to PNPO mutations: genotype, environment and treatment affect presentation and outcome, Brain 137 (2014) 1350–1360. [3] A. Ormazabal, M. Oppenheim, M. Serrano, A. Garcia-Cazorla, J. Campistol, A. Ribes, A. Ruiz, J. Moreno, K. Hyland, P. Clayton, S. Heales, R. Artuch, Pyridoxal 5′-phosphate values in cerebrospinal fluid: reference values and diagnosis of PNPO deficiency in paediatric patients, Mol. Genet. Metab. 94 (2008) 173–177. [4] E.J. Footitt, S.J. Heales, P.B. Mills, G.F. Allen, M. Oppenheim, P.T. Clayton, Pyridoxal 5′phosphate in cerebrospinal fluid; factors affecting concentration, J. Inherit. Metab. Dis. 34 (2011) 529–538. [5] M. van der Ham, M. Albersen, T.J. de Koning, G. Visser, A. Middendorp, M. Bosma, N.M. Verhoeven-Duif, M.G. de Sain-van der Velden, Quantification of vitamin B6 vitamers in human cerebrospinal fluid by ultra performance liquid chromatography-tandem mass spectrometry, Anal. Chim. Acta 712 (2012) 108–114. [6] M. Albersen, M. Bosma, J.J. Jans, F.C. Hofstede, P.M. van Hasselt, M.G. de Sain-van der Velden, G. Visser, N.M. Verhoeven-Duif, Vitamin B6 in plasma and cerebrospinal fluid of children, PLoS One 10 (2015), e0120972. . [7] D. Talwar, T. Quasim, D.C. McMillan, J. Kinsella, C. Williamson, D.S. O'Reilly, Optimisation and validation of a sensitive high-performance liquid chromatography assay for routine measurement of pyridoxal 5-phosphate in human plasma and red cells using pre-column semicarbazide derivatization, J. Chromatogr. B Anal. Technol. Biomed. Life Sci. 792 (2003) 333–343. [8] R. Spector, Vitamin B6 transport in the central nervous system: in vitro studies, J. Neurochem. 30 (1978) 889–897. [9] D. Kobayashi, T. Yoshimura, A. Johno, M. Ishikawa, K. Sasaki, K. Wada, Decrease in pyridoxal-5′-phosphate concentration and increase in pyridoxal concentration in rat plasma by 4′-O-methylpyridoxine administration, Nutr. Res. 35 (2015) 637–642.
Contents lists available at ScienceDirect
Clinica Chimica Acta journal homepage: www.elsevier.com/locate/clinchim
Measurement of pyridoxal 5′-phosphate, pyridoxal, and 4-pyridoxic acid in the cerebrospinal fluid of children Tomoyuki Akiyama a,⁎, Mari Akiyama a, Yumiko Hayashi a, Takashi Shibata a,b, Yoshiyuki Hanaoka a,b, Soichiro Toda c, Katsumi Imai d, Shin-ichiro Hamano e, Tohru Okanishi f, Harumi Yoshinaga a,b, Katsuhiro Kobayashi a,b a
Department of Child Neurology, Okayama University Hospital, Okayama, Japan Department of Child Neurology, Okayama University Graduate School of Medicine, Dentistry and Pharmaceutical Sciences, Okayama, Japan c Department of Pediatrics, Kameda Medical Center, Chiba, Japan d Department of Pediatrics, NHO Shizuoka Institute of Epilepsy and Neurological Disorders, Shizuoka, Japan e Division of Neurology, Saitama Children's Medical Center, Saitama, Japan f Department of Child Neurology, Seirei Hamamatsu General Hospital, Hamamatsu, Japan b
a r t i c l e
i n f o
Article history: Received 6 November 2016 Received in revised form 24 December 2016 Accepted 27 December 2016 Available online 28 December 2016 Keywords: Pyridoxal phosphate 4-pyridoxic acid Vitamin B6 Liquid chromatography Fluorescence detection Epilepsy
a b s t r a c t Background: We quantified pyridoxal 5′-phosphate (PLP), pyridoxal (PL), and 4-pyridoxic acid (PA) in the cerebrospinal fluid (CSF) of children and to investigate the effect of age, sex, epilepsy, and anti-epileptic drug (AED) therapy on these vitamers. Methods: CSF samples prospectively collected from 116 pediatric patients were analyzed. PLP, PL, and PA were measured using high-performance liquid chromatography with fluorescence detection, using pre-column derivatization by semicarbazide. Effects of age, sex, epilepsy, and AEDs on these vitamers and the PLP/PL ratio were evaluated using multiple linear regression models. Results: The PLP, PL, and PA concentrations were correlated negatively with age and the PLP/PL ratio was correlated positively with age. Multiple regression analysis revealed that the presence of epilepsy was associated with lower PLP concentrations and PLP/PL ratios but sex and AED therapy had no influence on these values. The observed ranges of these vitamers in epileptic and non-epileptic patients were demonstrated. Conclusions: We showed the age dependence of PLP and PL in CSF from pediatric patients. Epileptic patients had lower PLP concentrations and PLP/PL ratios than non-epileptic patients, but it is unknown whether this is the cause, or a result, of epilepsy. © 2016 Elsevier B.V. All rights reserved.
1. Introduction Vitamin B6 consists of various vitamers, including pyridoxal (PL), pyridoxamine, pyridoxine, their phosphorylated forms (pyridoxal 5′phosphate [PLP], pyridoxamine 5′-phosphate, and pyridoxine 5′-phosphate), and 4-pyridoxic acid (PA). PLP, a biologically active form of vitamin B6, is involved in numerous biochemical reactions as a cofactor [1]. PLP is an essential cofactor of glutamate decarboxylase, which converts glutamic acid, an excitatory neurotransmitter, to γ-aminobutyric acid, a main inhibitory neurotransmitter in the central nervous system. Lack of PLP results in intractable epileptic seizures, such as neonatal epileptic encephalopathy caused by pyridox[am]ine-phosphate oxidase (PNPO)
Abbreviations: AED, anti-epileptic drugs; PA, 4-pyridoxic acid; PDXK, pyridoxal kinase; PL, pyridoxal; PLP, pyridoxal 5′-phosphate; PNPO, pyridox[am]ine-phosphate oxidase. ⁎ Corresponding author at: Department of Child Neurology, Okayama University Hospital, 2-5-1 Shikata-cho, Kita-ku, Okayama, 700-8558, Japan. E-mail address: [email protected] (T. Akiyama).
http://dx.doi.org/10.1016/j.cca.2016.12.027 0009-8981/© 2016 Elsevier B.V. All rights reserved.
deficiency [2]. Insufficient PLP also affects monoamine neurotransmitter metabolism because aromatic L-amino acid decarboxylase, an enzyme to synthesize dopamine and serotonin from their precursors, is dependent on PLP. There are few reports on vitamin B6 in cerebrospinal fluid (CSF), especially for children. Reference values of PLP in children were reported using high-performance liquid chromatography (HPLC) with fluorescence detection [3,4]. Other vitamers have been measured using liquid chromatography-tandem mass spectrometry, which showed that PLP and PL are the main forms of vitamin B6 in the CSF [5,6]. Reference values for CSF PL and PA in children 1 year of age or older were reported in only one study [6]. The impact of some factors, such as sex, epilepsy, and anti-epileptic drugs (AEDs), on these vitamer concentrations needs to be taken into account to interpret the assay results correctly. The negative effect of AEDs on the CSF PLP and PL concentration, and lower PL concentrations in males have been reported in children [6]. However, another study reported no effect of AEDs or seizures on CSF PLP concentrations [4].
2
T. Akiyama et al. / Clinica Chimica Acta 466 (2017) 1–5
In this study, we quantified the PLP, PL, and PA concentrations in CSF obtained from pediatric patients to investigate the effect of age, sex, epilepsy, and AEDs on the concentrations of these vitamers and the PLP/PL ratio. We also compared these values with reference values in the literature. 2. Patients and methods 2.1. Patients CSF samples were prospectively collected between October 2014 and November 2016 from pediatric patients (1 month to 17 y) who underwent lumbar puncture to investigate neurological symptoms (e.g. epileptic seizures, developmental delay) or suspected meningitis. We excluded patients with: 1) vitamin B6 supplementation (highdose therapy or supplement intake other than regular food); 2) CSF samples not protected from light within 15 min; 3) metabolic diseases that may affect vitamin B6 metabolism; 4) acute neurological process such as acute encephalitis/encephalopathy, febrile seizures, and status epilepticus; 5) significant blood contamination; and 6) statistical outliers. This study was approved by the ethics committee of Okayama University Hospital. Written informed consent was obtained from the parents or guardians of all the patients before the procedure.
The HPLC system consisted of a Waters Alliance 2695 module with a Waters 2475 multi λ fluorescence detector (Waters). The mobile phase A was 60 mmol/l sodium phosphate buffer with EDTA-2Na at an approximate pH of 6.5 (18 mmol/l disodium hydrogen phosphate, 42 mmol/l sodium dihydrogen phosphate, and 0.4 g/l EDTA-2Na), the mobile phase B was 100% acetonitrile, and the mobile phase C was ultrapure water. A gradient mode was used (Table 1), and the injection volume was 30 μl. The separation was performed at a flow rate of 1.2 ml/min through a reverse-phase column (Atlantis T3, 3 μm, 3.0 × 50 mm, Waters Japan) at 35 °C with a guard column (Atlantis T3 VanGuard Cartridge, 3 μm, 3.9 × 5 mm, Waters). The PLP and PL derivatives were detected using excitation (Ex) at 370 nm and emission (Em) at 460 nm. PA was detected by its natural fluorescence (Ex 320 nm, Em 420 nm). Concentrations were calculated using area under the curve, with a single-point calibrator at 1024 nmol/l. The entire analysis time was 20 min.
2.4. Statistical analysis Statistical analysis was performed using JMP 4.0 (SAS Institute). Correlation was evaluated using Spearman's correlation coefficient. The independent effects of age, sex, epilepsy, and AEDs were tested using multiple regression analysis. The significance level was set to 0.05.
2.2. CSF collection protocol Because no rostrocaudal gradient of B6 vitamers has been reported [5], we used any fraction of the CSF samples in this study. Collected CSF samples were protected from light within 15 min and frozen within 2 h. The stability of CSF B6 vitamers for 10 h and plasma PLP for 12 h at room temperature was reported when the samples were kept in the dark [5,7]. They were stored at −80 °C until analysis. When the samples were collected outside Okayama University Hospital, they were shipped on dry ice to our laboratory. 2.3. Measurement of PLP, PL, and PA We obtained PLP monohydrate (82870), PL hydrochloride (P9130), PA (P9630), and ethylenediaminetetraacetic acid disodium salt (EDTA-2Na) dihydrate (09-1420) from Sigma Aldrich (St. Louis, MO, USA), sodium dihydrogen phosphate (197-09705), disodium hydrogen phosphate (197-02865), semicarbazide hydrochloride (192-00372), and glycine (073-00732) from Wako Pure Chemical Industries, Ltd. (Osaka, Japan). Acetonitrile (1.00030.2500) was from Merck Millipore (Billerica, MA, USA). Ultrapure water was prepared using a Direct-Q UV3 system (Merck Millipore). Simultaneous measurement of PLP, PL, and PA was conducted according to Talwar et al. [7] with some modifications. Primary stock solutions of PLP, PL, and PA were prepared, mixed, and diluted with ultrapure water to achieve final concentrations of 250 μmol/l for each component. This secondary stock solution was further diluted with ultrapure water to 1024 nmol/l to make a single-point calibrator. Quality control samples at 64 nmol/l were also prepared. Derivatization of samples was conducted as described below. Briefly, 200 μl of undiluted CSF samples or 100 μl of serum/plasma samples diluted two-fold with 100 μl ultrapure water were derivatized by adding 16 μl of derivatization agent (370 mg/ml of semicarbazide hydrochloride and 250 mg/ml of glycine dissolved in ultrapure water). After vortexing and incubating the samples at 40 °C in the dark for 30 min, 16 μl of 70% perchloric acid was added and vortexed for deproteinization. After centrifuging for 5 min at 16,000 × g, 170 μl of supernatant was transferred to a new microtube and 23 μl of 25% sodium hydroxide was added to adjust the pH to 3.0–5.0. The samples were then filtered through 0.45-μm Millex LH filters (Merck Millipore) and subjected to HPLC.
3. Results 3.1. Patient characteristics There were 178 patients who underwent lumbar puncture to measure PLP, PL, and PA concentrations in the CSF. Sixty-two patients were excluded for the following reasons: 1) 40 patients with CSF samples protected from light N 15 min after collection; 2) 9 patients with acute neurological process; 3) 8 patients with metabolic diseases; 4) 2 patients on high-dose PLP therapy; and 5) 3 patients considered to be statistical outliers, whose PLP, PL, and PA concentrations were all above the cut-off values (3rd quartile + 1.5 × interquartile range). The data from the remaining 116 patients (68 males and 48 females) were used for analysis in this study. Patient age ranged from 1 to 199 months (median age, 48 months). There were 62 patients with a diagnosis of epilepsy (median age, 32.5 months), 50 of whom were taking AEDs. There were 54 patients without epilepsy (median age, 75.5 months), 6 of whom were taking AEDs. For these 6 patients, AEDs were administered for suspected epileptic seizures, or were used as muscle relaxants for dystonia or spasticity, or as mood stabilizers. In total, 56 patients were taking AEDs (median age, 45.5 months) and 60 were not taking AEDs (median age, 49.5 months). The administered AEDs were valproic acid in 29 patients, levetiracetam in 14 patients, zonisamide in 13 patients, phenobarbital in 9 patients, carbamazepine in 8 patients, topiramate in 8 patients, clonazepam in 8 patients, lamotrigine in 7 patients, clobazam in 7 patients, ethosuximide in 3 patients, potassium bromide in 2 patients, nitrazepam in 1 patient, and rufinamide in 1 patient either as monotherapy or in combination. Table 1 High-performance liquid chromatography gradient settings. Time (min)
Flow (ml/min)
Mobile phase A (%)
Mobile phase B (%)
Mobile phase C (%)
Gradient curve
Initial 3.5 7.0 7.1 10.0 20.0
1.2 1.2 1.2 1.2 1.2 1.2
98.6 98.6 94.0 0.0 98.6 98.6
1.4 1.4 6.0 60.0 1.4 1.4
0.0 0.0 0.0 40.0 0.0 0.0
Linear Linear Step Step Linear
T. Akiyama et al. / Clinica Chimica Acta 466 (2017) 1–5
3.2. Validity of PLP, PL, and PA assay using HPLC A representative chromatogram of a standard sample and a CSF sample is shown in Fig. 1. PLP, PA, and PL were eluted at 2.3 min, 3.7 min, and 7.2 min, respectively. The limit of detection (signal-to-noise [S/N] ratio ≥ 3) was 1.5 nmol/l for PLP, 1.0 nmol/l for PL, and 0.4 nmol/l for PA. The limit of quantification (LOQ, S/N ratio ≥ 10) was 3.5 nmol/l for PLP, 2.0 nmol/l for PL, and 1.2 nmol/l for PA. There was a good linear relationship over the concentration range of LOQ to 1024 nmol/l for PLP (R2 = 0.9996), PL (R2 = 0.9999), and PA (R2 N 0.9999). Because the regression line intercepts for these compounds were not significantly different from zero, we chose to use a single-point calibrator at 1024 nmol/l. The intraday coefficients of variation (CV) were 0.9% for PLP, 0.3% for PL, and 0.2% for PA (n = 5) for a quality control sample at 64 nmol/l. Inter-day CVs were 1.1% for PLP, 1.4% for PL, and 0.9% for PA (n = 5). PLP, PL, and PA recovery from the CSF samples spiked with 25.6 nmol/l (n = 3) was 95.7– 98.8% (mean, 96.8%), 100.1–101.4% (mean, 100.8%), and 96.6–98.0% (mean, 97.1%), respectively. PLP, PL, and PA recovery from the CSF samples spiked with 102.4 nmol/l (n = 3) was 98.5–100.2% (mean, 99.2%), 100.0– 100.1% (mean, 100.1%), and 99.3–101.7% (mean, 100.2%), respectively.
3
incorporating age, sex, presence of epilepsy, and AED therapy showed that presence of epilepsy (p = 0.0015) and age (p b 0.0001) were associated with lower PLP concentrations (Table 2). Sex and AED therapy did not affect the PLP concentration. The PL concentration was negatively correlated with age (ρ = −0.728, p b 0.0001; Fig. 2B). Multiple regression analysis showed that only age (p b 0.0001) influenced the PL concentration (Table 2). Sex, presence of epilepsy, and AED therapy did not affect the PL concentration. The PLP/PL ratio was positively correlated with age (ρ = 0.524, p b 0.0001; Fig. 2C). Multiple regression analysis demonstrated that age (p b 0.0001) and presence of epilepsy (p = 0.0427) independently influenced the PLP/PL ratio (Table 2). Sex and AED therapy did not affect the PLP/PL ratio. We did not perform a statistical analysis on PA, because most patients except for young patients had PA concentrations below the LOQ (b1.2 nmol/l; Fig. 2D). The observed ranges of PLP, PL, PA, and PLP/PL according to age groups and presence/absence of epilepsy are presented in the Table 3. 4. Discussion
3.3. PLP, PL, and PA concentrations, and the PLP/PL ratio The PLP concentration was negatively correlated with age (ρ = − 0.310, p = 0.0007; Fig. 2A). Multiple regression analysis
We quantified the PLP, PL, and PA concentrations, and the PLP/PL ratio in pediatric CSF samples and showed their dependence on age. Although sex and AED therapy did not affect the concentrations of these
Fig. 1. Chromatograms of a standard sample and a CSF sample. A) A standard sample of PLP, PL, and PA at 1024 nmol/l. B) A CSF sample from a 7-month-old patient. PLP, PL, and PA were well separated from other peaks (PLP, 28.1 nmol/l; PL, 52.9 nmol/l; and PA, 1.5 nmol/l). PLP, pyridoxal 5′-phosphate; PL, pyridoxal; PA, 4-pyridoxic acid; Ex., excitation; Em., emission.
4
T. Akiyama et al. / Clinica Chimica Acta 466 (2017) 1–5
Fig. 2. PLP, PL, and PA concentration and PLP/PL ratio in the cerebrospinal fluid according to age. A) PLP vs. age. PLP was correlated negatively with age. B) PL vs. age. PL was correlated negatively with age. C) PLP/PL ratio vs. age. PLP/PL ratio was positively correlated with age. D) PA vs. age. PA showed high concentrations at young ages. Filled circles, non-epileptic patients; open squares, epileptic patients; PLP, pyridoxal 5′-phosphate; PL, pyridoxal; PA, 4-pyridoxic acid.
vitamers, epileptic patients had lower PLP values and PLP/PL ratios than non-epileptic patients. The dependence of PLP on age (higher PLP values at a younger age) is consistent with past studies [3,4]. Our study showed a similar age
Table 2 Results of multiple regression analysis. Vitamers
Variables
p
PLP
Age Sex Epilepsy AED Age Sex Epilepsy AED Age Sex Epilepsy AED
b0.0001⁎ NS 0.0015⁎⁎
PL
PLP/PL
NS b0.0001⁎ NS NS NS b0.0001⁎⁎⁎ NS 0.0427⁎⁎ NS
PLP, pyridoxal 5′-phosphate; PL, pyridoxal; AED, anti-epileptic drugs; NS, not significant. ⁎ Negative correlation with age. ⁎⁎ Presence of epilepsy was associated with lower values. ⁎⁎⁎ Positive correlation with age.
dependence for PL concentration, which was not investigated in a previous study that measured CSF PL in children [6]. The degree of decline with age seems to be different between PLP and PL based on the age dependence of the PLP/PL ratio. The observed PLP, PL, and PA ranges and the PLP/PL ratio in non-epileptic patients were similar to previously reported values [3,4,6]. There is not as much information on the PL assay as on the PLP assay, and measurement of both PL and PLP may provide useful clinical information. Production of PLP in the central nervous system requires an appropriate supply of non-phosphorylated vitamin B6 vitamers, including PL, from the peripheral blood and sufficient activity of pyridoxal kinase (PDXK) and PNPO [1]. Because PLP is not membrane-permeable, PLP in the blood has to be dephosphorylated to PL by a membrane-bound tissue non-specific alkaline phosphatase [1]. After PL crosses the cell membrane from the blood, it is phosphorylated back to PLP by PDXK and trapped within the cell. Although the brain itself does not release PLP, the choroid plexus releases PLP into the CSF [8]. When CSF PLP is low, the PL assay may clarify whether it is caused by deficient phosphorylation. Some toxins (e.g. ginkgotoxin) and medications (e.g., isoniazid) are known to inhibit PDXK [9]. Association between CSF PLP values and AED therapy or epilepsy has not been consistently shown. Footitt et al. demonstrated no statistical difference of CSF PLP concentrations between patients with and without
T. Akiyama et al. / Clinica Chimica Acta 466 (2017) 1–5 Table 3 Observed ranges of PLP, PL, PA, and PLP/PL according to age. PLP (nmol/l) Patients without epilepsy 1–11 months 16.9–51.4 (n = 12) (median 37.2) 1–6 years (n = 15) 15.3–49.3 (24.5) 7–17 years (n = 27) 13.1–41.2 (22.6) Patients with epilepsy 1–11 months (n = 22) 1–6 years (n = 29) 7–17 years (n = 11)
5.9–56.3 (30.9) 8.7–50.8 (18.9) 6.6–32.0 (17.6)
PL (nmol/l)
PA (nmol/l)
PLP/PL
17.5–97.5 (51.6) 11.7–34.8 (26.4) 12.0–38.9 (17.8)
b1.2a–3.3 (1.4) b1.2a–1.3 (b1.2a) b1.2a
0.24–0.99 (0.53) 0.76–1.77 (1.08) 0.58–2.25 (1.10)
21.2–99.8 (47.4) 7.1–44.4 (25.0) 7.9–25.0 (16.3)
a
b1.2 –3.2 (1.3) b1.2a–2.1 (b1.2a) b1.2a
0.14–1.47 (0.56) 0.39–1.54 (0.75) 0.83–1.96 (1.01)
PLP, pyridoxal 5′-phosphate; PL, pyridoxal; PA, 4-pyridoxic acid. a Below the limit of quantification (1.2 nmol/l).
seizures, and between those with and without AED therapy [4]. Albersen et al. showed that PLP and PL concentrations were lower in patients with AEDs than in those not receiving AEDs but these values were not influenced by epilepsy [6]. The current study showed that there is a lower PLP concentration and PLP/PL ratio in epileptic patients. This effect was significant even when AED use is taken into account. A negative impact of epilepsy on the PLP concentration and PLP/PL ratio suggests that decreased PLP production may play a role in the pathophysiology of epilepsy or that low PLP values are the result of epilepsy, but it is difficult to distinguish between these two potential mechanisms at this point. Although the use of AEDs was not associated with PLP and PL in the current study, we did not have enough data to investigate the effect of each kind of AED. The majority of epileptic patients were treated with AEDs at the time of lumbar puncture, which makes it difficult to separate the effect of AED from that of epilepsy. We cannot fully exclude the possibility that certain AEDs that are used mainly for epileptic patients in this study had an effect on PLP and PL concentrations in the CSF. For PLP and PL measurement, we used pre-column derivatization by semicarbazide hydrochloride, which forms stable and strongly fluorescent derivatives, thereby making sample handling easy and enhancing the assay sensitivity. This derivatization process was reported to be complete within 30 min at the ambient temperature [7]. We chose to perform this reaction at 40 °C because we experienced lower and variable recoveries (80–95%) of PLP and PL with undiluted serum/plasma samples derivatized at the ambient temperature. This study has some limitations. A larger number of samples will clarify whether epilepsy or certain kinds of AEDs influence the CSF PLP concentration. Although we excluded patients with possible alterations in vitamin B6 metabolism, some patients may have unknown disorders that can affect the assay results. Because CSF PL is influenced by plasma PL [6], further study with simultaneous measurement of B6 vitamers in the CSF and plasma is necessary to understand the overall dynamics of these vitamers. We are currently in the process of collecting paired CSF and plasma samples for this purpose. 5. Conclusion We quantified the PLP, PL, and PA concentration in the CSF of pediatric patients with neurological symptoms. The PLP, PL, and PA
5
concentrations decreased with age but the PLP/PL ratio increased with the age. Patients with epilepsy had lower CSF PLP concentrations and PLP/PL ratios than those without epilepsy, which might be either the cause or a result of epilepsy. Use of AEDs did not affect these values. Observed ranges of PLP, PL, and PA concentrations were also demonstrated. Financial disclosure This study was supported by the Japan Society for the Promotion of Science (Grant Number JP15K09622). This funding source had no involvement in the study design, the collection, analysis, and interpretation of data, the writing of the report, or the decision to submit this article. Acknowledgements T.A. received grant support from the Japan Society for the Promotion of Science. We thank Dr. Shinpei Baba at Tokyo Medical and Dental University, Dr. Tetsuhiro Fukuyama at Nagano Children's Hospital, Dr. Tomohide Goto at Kanagawa Children's Medical Center, Dr. Yoshikazu Kitami at Tokyo Metropolitan Children's Medical Center, Dr. Masaya Kubota at National Center for Child Health and Development, Ichiro Kuki at Osaka City General Hospital, Dr. Satoko Kumada at Tokyo Metropolitan Neurological Hospital, Dr. Shinji Saito at Nagoya City University, Dr. Takashi Shiihara at Gunma Children's Medical Center, Dr. Kentaro Shirai at Tsuchiura Kyodo Hospital, Dr. Akihito Takahashi at Kurashiki Central Hospital, and Dr. Yoshihiro Toda at Tokushima University, for providing CSF samples from pediatric patients. We thank Eibunkousei. net (http://www.eibunkousei.net/) for English-language editing. References [1] R. Surtees, P. Mills, P. Clayton, Inborn errors affecting vitamin B6 metabolism, Future Neurol. 1 (2006) 615–620. [2] P.B. Mills, S.S. Camuzeaux, E.J. Footitt, K.A. Mills, P. Gissen, L. Fisher, K.B. Das, S.M. Varadkar, S. Zuberi, R. McWilliam, T. Stodberg, B. Plecko, M.R. Baumgartner, O. Maier, S. Calvert, K. Riney, N.I. Wolf, J.H. Livingston, P. Bala, C.F. Morel, F. Feillet, F. Raimondi, E. Del Giudice, W.K. Chong, M. Pitt, P.T. Clayton, Epilepsy due to PNPO mutations: genotype, environment and treatment affect presentation and outcome, Brain 137 (2014) 1350–1360. [3] A. Ormazabal, M. Oppenheim, M. Serrano, A. Garcia-Cazorla, J. Campistol, A. Ribes, A. Ruiz, J. Moreno, K. Hyland, P. Clayton, S. Heales, R. Artuch, Pyridoxal 5′-phosphate values in cerebrospinal fluid: reference values and diagnosis of PNPO deficiency in paediatric patients, Mol. Genet. Metab. 94 (2008) 173–177. [4] E.J. Footitt, S.J. Heales, P.B. Mills, G.F. Allen, M. Oppenheim, P.T. Clayton, Pyridoxal 5′phosphate in cerebrospinal fluid; factors affecting concentration, J. Inherit. Metab. Dis. 34 (2011) 529–538. [5] M. van der Ham, M. Albersen, T.J. de Koning, G. Visser, A. Middendorp, M. Bosma, N.M. Verhoeven-Duif, M.G. de Sain-van der Velden, Quantification of vitamin B6 vitamers in human cerebrospinal fluid by ultra performance liquid chromatography-tandem mass spectrometry, Anal. Chim. Acta 712 (2012) 108–114. [6] M. Albersen, M. Bosma, J.J. Jans, F.C. Hofstede, P.M. van Hasselt, M.G. de Sain-van der Velden, G. Visser, N.M. Verhoeven-Duif, Vitamin B6 in plasma and cerebrospinal fluid of children, PLoS One 10 (2015), e0120972. . [7] D. Talwar, T. Quasim, D.C. McMillan, J. Kinsella, C. Williamson, D.S. O'Reilly, Optimisation and validation of a sensitive high-performance liquid chromatography assay for routine measurement of pyridoxal 5-phosphate in human plasma and red cells using pre-column semicarbazide derivatization, J. Chromatogr. B Anal. Technol. Biomed. Life Sci. 792 (2003) 333–343. [8] R. Spector, Vitamin B6 transport in the central nervous system: in vitro studies, J. Neurochem. 30 (1978) 889–897. [9] D. Kobayashi, T. Yoshimura, A. Johno, M. Ishikawa, K. Sasaki, K. Wada, Decrease in pyridoxal-5′-phosphate concentration and increase in pyridoxal concentration in rat plasma by 4′-O-methylpyridoxine administration, Nutr. Res. 35 (2015) 637–642.