Article Text

Download PDFPDF

Short report
Confirmation of mutations in PROSC as a novel cause of vitamin B 6 -dependent epilepsy
  1. Barbara Plecko1,2,
  2. Markus Zweier3,
  3. Anaïs Begemann2,3,
  4. Deborah Mathis4,
  5. Bernhard Schmitt1,
  6. Pasquale Striano5,
  7. Martina Baethmann6,
  8. Maria Stella Vari5,
  9. Francesca Beccaria7,
  10. Federico Zara8,
  11. Lisa M Crowther1,
  12. Pascal Joset3,
  13. Heinrich Sticht9,
  14. Sorina Mihaela Papuc2,3,
  15. Anita Rauch2,3
  1. 1 Division of Child Neurology, University Children’s Hospital Zurich, Zurich, Switzerland
  2. 2 radiz – “Rare Disease Initiative Zurich, Clinical Research Priority Program for Rare Diseases University of Zurich”, University of Zurich, Zurich, Switzerland
  3. 3 Institute of Medical Genetics, University of Zurich, Zurich, Switzerland
  4. 4 Clinical Chemistry, University Children’s Hospital Zurich, Zurich, Switzerland
  5. 5 Pediatric Neurology and Muscular Diseases Unit, Department of Neurosciences, Rehabilitation, Ophthalmology, Genetics, Maternal and Child Health, University of Genova, G. Gaslini Institute, Genova, Italy
  6. 6 Sozialpädiatrisches Zentrum, Dritter Orden, München-Nymphenburg, Munich, Germany
  7. 7 Child Neuropsychiatry, C. Poma Hospital, Mantova, Italy
  8. 8 IRCCS, University of Genova, G. Gaslini Institute, Genova, Italy
  9. 9 Institute of Biochemistry, University of Erlangen-Nuernberg, Erlangen, Germany
  1. Correspondence to Professor Barbara Plecko, Division of Child Neurology, University Children’s Hospital, Steinwiesstrasse 75, 8032 Zurich; Barbara.Plecko{at}


Vitamin-B6-dependent epilepsies are a heterogenous group of treatable disorders due to mutations in several genes (ALDH7A1, PNPO, ALPL or ALDH4A1). In neonatal seizures, defects in ALDH7A1 and PNPO explain a major fraction of cases. Very recently biallelic mutations in PROSC were shown to be a novel cause in five families. We identified four further unrelated patients harbouring a total of six different mutations, including four novel disease mutations. Vitamin B6 plasma profiles on pyridoxine did not enable the differentiation of patients with PROSC mutations. All four patients were normocephalic and had normal cranial imaging. Pyridoxine monotherapy allowed complete seizure control in one, while two patients had occasional febrile or afebrile seizures and one needed additional valproate therapy for photosensitive seizures. Two patients underwent a controlled pyridoxine withdrawal with signs of encephalopathy within a couple of days. Three had favourable outcome with normal intellectual properties at age 12.5, 15.5 and 30 years, respectively, while one child had marked developmental delay at age 27 months. The clinical and electroencephalographic phenotype in patients with PROSC mutations was indistinguishable from ALDH7A1 and PNPO deficiency. We therefore confirm PROSC as a novel gene for vitamin-B6-dependent epilepsy and delineate a non-specific plasma vitamin B6 profile under pyridoxine treatment.

  • neonatal seizures
  • inborn errors of metabolism
  • vitamin B6
  • pyridoxine
  • PROSCexome sequencing.

Statistics from


Very recently Darin et al 1 described homozygous or compound heterozygous mutations in PROSC in seven patients of five families with vitamin-B6-dependent epilepsy. Hitherto known biochemical mechanisms underlying vitamin-B6-dependent epilepsies were impaired formation or transport of pyridoxal 5’-phosphate (PLP) or its inactivation by small accumulating compounds.2 PROSC encodes a PLP-binding protein and mutations most probably lead to impaired cellular PLP homeostasis.1 PLP is an abundant cofactor of various enzymatic reactions in amino acid and neurotransmitter metabolism. Any mechanism leading to reduced PLP availability may therefore result in seizures and altered amino acid or neurotransmitter profiles.3 Six out of seven patients with PROSC mutations known so far presented with neonatal seizure onset within 24 hours and a burst suppression pattern in five out of seven. Four out of seven patients had a neonatal head circumference (HC) <10%, cranial imaging was abnormal in four out of seven and none of the patients had normal cognitive function. PLP concentrations in plasma and cerebrospinal fluid (CSF) were abnormal in all patients tested. Only one patient was seizure-free on pyridoxine monotherapy. Four were reported as having benefited from a switch to PLP, which was accompanied by a simultaneous dose increase in some.

We here present clinical and molecular data of four novel unrelated patients harbouring a total of six homozygous or compound heterozygous mutations in PROSC, confirming and further delineating this new entity causing vitamin-B6-dependent epilepsy.


All patients had normal sequencing of ALDH7A1 and PNPO by Sanger sequencing. Patient (P) samples and questionnaires were referred after full informed consent. This study was approved by the local ethical committee. DNA was extracted from peripheral blood. In P1 and P2 and their healthy parents trio whole exome sequencing (WES) was performed using the Agilent (Santa Clara, California, USA) SureSelect XT Clinical Research Exome Kit (V5) on a HiSeq2500 System (Illumina, San Diego, California, USA). The WES data were aligned and analysed for rare (minor allele frequency below 2%) non-synonymous exonic and splice site (including 12 intronic bp) variants considering both dominant and recessive modes of inheritance using NextGENe Software (SoftGenetics, State College, Pennsylvania, USA). PROSC variants were confirmed by Sanger sequencing in the patients and parents and were denominated according to RefSeq accession number NM_007198.3. Variants in P3 and P4 were identified in an additional PROSC screen by Sanger sequencing in a collection of nine patients with clear vitamin-B6-dependent epilepsy but unclear genetic background. Inheritance pattern of the variants was confirmed by Sanger sequencing in the parents. Mutation modelling was performed by standard procedures4 using the structure of a homologous yeast protein (PDB code: 1CT5) as a template. Vitamin B6 vitamers were analysed in plasma of all four patients as described5 and a kinetic 6 hours plasma profile was performed in P1 after the oral intake of 300 mg of pyridoxine as a single dose. P1, P2 and P3 had analysis of plasma amino acids and P2 also of amino acids, neurotransmitters and 5-methyltetrahydrofolate (5-MTHF) as well as total and free GABA in CSF. P1 had testing for alpha aminoadipic semialdehyde (AASA) and P1 and P2 of pipecolic acid (PA) in plasma.


Full details are provided in table 1.

Table 1

Clinical and molecular data

Patient 1

This female neonate presented with irritability and sleeplessness followed by seizures with grimacing, roving eye movements and tremor with no response to phenobarbitone on day 7. EEG showed mild slowing of background activity. She was transferred to a tertiary care centre where persistent seizure activity was documented on EEG. A first dose of pyridoxine, 100 mg i.v. (13 mg/kg) on day 28 immediately stopped seizures and was accompanied by muscular hypotonia, periodic breathing and continuous sleep over 2 days. She was seizure-free on pyridoxine 10 mg/kg/day p.o. Motor and speech development were normal. At age 14 months a controlled pyridoxine withdrawal lead to marked irritability and presumed hallucinations after 5–6 days off pyridoxine, followed by sleep disturbances, inconsolable crying and panic attacks until day 10, which all resolved after 100 mg of pyridoxine i.v. Since then she stayed seizure-free on pyridoxine monotherapy (300–400 mg/day p.o), but used to have hallucinations and panic attacks along febrile illnesses, overcome by transient doubling of the pyridoxine dose. At age 12 years she performs well at her sixth class in grammar school.

Patient 2

This girl presented from day 5 with poor feeding, irritability, sleeplessness and tremor followed by cloni of arms and legs on day 6 with desaturation and eye deviation lasting several minutes. The EEG showed discontinuity of background activity as well as multifocal spikes. A single dose of phenobarbitone stopped the seizures. 100 mg of pyridoxine i.v. given on day 7 had no impact on EEG but her feeding improved. From day 16 seizures recurred and phenobarbitone was administered regularly. On day 26 during a viral infection she had breakthrough seizures unresponsive to increased doses of phenobarbitone with prompt cessation following 100 mg of pyridoxine i.v. From then on she was continued on pyridoxine treatment, 100 mg/day p.o in two SD. Due to one unprovoked seizure at age 4 months and a total of five febrile seizures during infancy and early childhood, pyridoxine was increased to 25 mg/kg/day p.o. Motor milestones were slightly delayed, but speech development was normal. She stayed seizure-free on 250 mg of pyridoxine/day p.o. and EEGs were normal. A controlled pyridoxine withdrawal at age 12 years lead to abdominal discomfort, nausea, vertigo, headaches, paraesthesia and distal weakness, which resolved after reintroduction of pyridoxine. She had one similar attack at age 14 years preceding her last afebrile seizures while on pyridoxine 150 mg/day p.o. The EEG showed bilateral occipital slowing. She attends the nineth class of grammar school with good performance.

Patient 3

In this male patient tonic-clonic seizures started from day 3. Interictal EEG showed right frontotemporal discharges. Seizures were resistant to treatment with phenobarbitone, levetiracetam, valproic acid and ACTH. At age 1 month he first received 100 mg of pyridoxine i.v., which was well tolerated and clearly improved seizure frequency. Since then he experienced occasional tonic clonic seizures while on monotherapy with pyridoxine, 450 mg/day p.o. EEG showed occasional occipital discharges. His developmental milestones were markedly delayed. Independent sitting was achieved at age 20 months. At age 27 months he is not able to walk independently and speech development is absent.

Patient 4

He presented with tonic seizures and a burst suppression pattern on day 9. Seizures were resistant to treatment with phenobarbitone and nitrazepam, but responded to 100 mg of pyridoxine i.v. at 1 month of age. EEG changes resolved over 2 weeks. Due to the appearance of photosensitive tonic-clonic seizures, valproate (30 mg/kg/day) was introduced from age 6 years. Since then the patient went seizure-free. His EEGs were unremarkable. His development was good except for mild learning difficulties in regular school. At age 30 years he has a driving license and is working in a supermarket.


WES revealed compound heterozygosity for two very rare probably damaging missense variants in P1 (one of which is also reported in trans with another mutation by Darin et al in subject 7)1 and compound heterozygosity for a novel frameshift mutation and a very rare probably damaging missense mutation in P2. P3 was found to harbour a homozygous missense variant with ambiguous predictions, which was however already reported in subject 7.1 P4 was found to carry a homozygous novel missense variant with damaging predictions that may also affect splicing. All detected variants affected highly conserved amino acids within the PLP-binding barrel domain similar to the alanine racemase N terminal domain, and with reference to the specificity of the phenotype, were considered disease causing (table 2).

Table 2

Summary of PROSC mutations identified in this study

Results of vitamin B6 plasma profiles are provided in table 3.

Table 3

Plasma vitamin Bprofiles of patients with PROSC mutations on pyridoxine and ranges for vitamin B6 reference profiles of controls and patients with antiquitin, PNPO and TNSALP deficiency on pyridoxine5

Results of biochemical analyses are provided in table 4. Plasma aminoacids in P3 were determined before first application of pyridoxine and were normal. In P1, plasma amino acids were normal while on pyridoxine at age 11 years and 12 years, while P2 showed elevated alanine, proline and glycine at age 8 days, 24 hours after a single pyridoxine dosage of 100 mg i.v. and elevation of alanine and proline at age 6 weeks, 2 weeks after the initiation of continuous pyridoxine administration. Amino acids in CSF were determined in P2 along the plasma amino acids and showed 1.5–3-fold elevation of mean normal concentrations for alanine, valine, isoleucine, leucine, glycine and, inconsistently for tyrosine. Homocarnosine in CSF was not determined. Neurotransmitter analysis, determination of 5-MTHF and total and free GABA gave normal results in the CSF sample of P2 taken 2 weeks after the initiation of continuous pyridoxine administration.

Table 4

Biochemical data for patients with PROSC mutations


We report on four novel unrelated patients harbouring four novel and two known homozygous or compound heterozygous mutations in PROSC. While the first report on patients with PROSC mutations indicated low HC at birth as a common sign, all of our patients were normocephalic at birth and during long-term follow-up. Seizure onset in our patients was somewhat later than previously reported and occurred between days 3 and 9, and only one out of four had a burst suppression pattern. Comparable to Darin et al,1 only one patient had complete seizure control on pyridoxine monotherapy, while two patients had occasional afebrile seizures and one needed additional valproate therapy for photosensitive seizures. None of our patients had been switched to PLP, but this was intended for patient 3 after the diagnosis of PROSC deficiency was established. While four out of seven patients described by Darin et al 1 had underdeveloped white matter or broad sulci, all our patients had normal MRI. In P1, cMRI had been performed on day 12 only but was not repeated due to normal development. All four patients presented here had markedly elevated plasma PLP concentrations while on pyridoxine. Compared with a robust control group,5 only two of these patients had values above those seen in patients with ALDH7A1 or PNPO deficiency and in patients with TNSALP deficiency (congenital hypophosphatasia) on pyridoxine.5 6 Thus the vitamin B6 plasma profile is not suitable for a clear differentiation of patients with PROSC mutations while on treatment with vitamin B6. In contrast to the observation by Darin et al,1 we could show on the basis of the kinetic study performed in P1, that pyridoxine is detectable in plasma of patients with PROSC mutations depending on the interval to last pyridoxine intake. While all other defects causing vitamin-B6-dependent epilepsies can be detected by their respective biomarkers irrespective of treatment with vitamin B6,5 7 patients with PROSC mutations hitherto lack a specific biomarker, as low pretreatment PLP concentrations in CSF have been seen with other inborn errors causing vitamin-B6-dependent epilepsies.8 In accordance with two previously published patients with PROSC mutations who had analysis of CSF amino acid profiles,1 we found elevation of several amino acids that are transaminated or degraded by vitamin-B6-dependent enzymes in one patient (P2) even 2 weeks after continuous administration of pyridoxine, 100 mg/day in two SD These findings are unspecific, as amino acid (and neurotransmitter) changes have been described in other inborn errors with vitamin-B6-dependent epilepsies,3 but are important diagnostic indicators of disturbed vitamin B6 metabolism. As homocarnosine was not measured in the two CSF samples of P2 we unfortunately cannot comment on its potential role as a specific biomarker for PROSC deficiency as suggested by Darin et al.1 In contrast, plasma amino acids were normal in one neonate (P3) before the first administration of pyridoxine, while P2 had elevated glycine and proline in the sample taken 24 hours after a single first pyridoxine dosage of 100 mg i.v. These findings illustrate that, in accordance to other inborn errors leading to vitamin-B6-dependent epilepsies, also in PROSC mutations the impact of reduced PLP availability is more pronounced in the central nervous system than in other organ systems.

In contrat to the seven cases reported previously,1 three out of four patients with PROSC mutations reported here had normal outcome at age 12.5, 15.5 and 30 years, including one with mild learning difficulties while one patient (P3) had marked developmental delay. Notably P3 is homozygous for the c.260C>T (p.Pro87Leu) mutation, which was found compound heterozygous with c.722G>A (p.Arg241Gln) in subject 71 with an attenuated phenotype. Therefore, the milder disease course attributed to the c.260C>T mutation is not supported by our observation. Instead, subject 7’s1 milder disease course might be explained by the c.722G>A mutation, which is also present in our P1 with normal intellectual outcome at age 12 years and complete seizure freedom on pyridoxine monotherapy. Since all the other mutations observed in our patients are novel, the overall milder phenotype in patients P1, P2 and P4 in comparison to the previously described patients might be explained by less severe functional impacts of these novel mutations.

In summary, our findings further evidence PROSC as a novel, presumably relatively common gene causing vitamin-B6-dependent seizures and demonstrate that plasma vitamin B6 profiles while on pyridoxine are not suitable for the detection of this new entity.


The authors thank Céline Bürer for the PNPO and ALDH7A1 mutation analysis.



  • Contributors BP set up the study design, contributed biological samples of patients and drafted the manuscript.

    DM has established the method for vitamin B6 vitamer analysis, ran and interpreted all vitamin B6 assays and gave substantial input to the manuscript.

    BS has contributed biological samples of patients and critically reviewed the manuscript.

    MB provided biological samples and critically reviewed the manuscript.

    PS, MSV and FB provided clinical data and biological samples of patients and revised the manuscript.

    FZ screened some patients for PNPO and ALDH7A1 mutations and provided DNA samples.

    LC has helped in data interpretation and critically reviewed the manuscript.

    MZ and AR contributed substantially to the conception and design of the study, supervised the genetic data acquisition and interpretation and gave major input to the manuscript.

    AB performed the Sanger sequencing, contributed to interpretation of variants and gave input to the manuscript.

    SMP and PJ contributed to the acquisition and interpretation of the whole exome sequencing data.

    HS contributed the mutational modelling.

  • Competing interests None declared.

  • Ethics approval IRB of Zurich.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Correction notice This article has been updated since it published Online First. The author name Maria Stella Vari has been corrected.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.