Address correspondence to Orly Elpeleg, Department of Genetic and Metabolic Diseases, Hadassah, Hebrew University Medical Center, Jerusalem 91120, Israel. E-mail: firstname.lastname@example.org
West syndrome consists of infantile spasms, hypsarrhythmia, and developmental arrest. Most patients remain mentally retarded and many develop Lennox-Gastaut syndrome. Using homozygosity mapping followed by exome sequencing we identified an ST3GAL3 mutation in three infants with West syndrome. ST3GAL3 encodes a sialyltransferase involved in the biosynthesis of sialyl-Lewis epitopes on cell surface–expressed glycoproteins. The mutation affected an essential sialyl-motif and abolished enzymatic activity. Abnormalities in proteins involved in forebrain γ-aminobutyric acid (GABA)ergic synaptic growth and function were recently proposed to account for infantile spasms. Dysfunctional ST3GAL3 may thus result in perturbation of the posttranslational sialylation of proteins in these pathways.
West syndrome (WS) is an age-dependent epileptic encephalopathic syndrome defined by a triad of infantile spasms, hypsarrhythmic pattern on electroencephalography (EEG) and developmental arrest or regression (Kato, 2006). The prognosis is generally grave, with 70–90% of the patients remaining mentally retarded and 20–50% developing Lennox-Gastaut syndrome (Trevathan et al., 1999). WS is classified according to the underlying etiology into (1) an acquired WS, (2) a congenital/developmental WS, and (3) WS of unknown etiology (Berg et al., 2010). The acquired cases include sequelae of central nervous system (CNS) infections, hypoxic ischemic, and posttraumatic injuries, whereas the congenital/developmental WS are genetic, resulting mostly from mutations in single genes that share functional links. By now mutations in ARX, CDKL5, STXBP1, SLC25A22, and SPTAN1 (reviewed by Paciorkowski et al., 2011) have already been linked to WS, and many more are yet to come with the introduction of high throughput genetic sequencing.
We describe the molecular investigation of four infants from a consanguineous family who presented with WS evolving to Lennox-Gastaut syndrome with severe intellectual disability.
Patients and Methods
The subjects of this study were three infants, two male and one female (Patients 4694, 4695, and 4696), the offspring of consanguineous parents of Palestinian origin, and their second degree cousin, 4713 (Fig. 1A). The patients shared a common phenotype (summarized in the Table 1) consisting of infantile spasms of the flexor type mainly, appearing at 3–7 months of age. These were accompanied by a hypsarrhythmic pattern on EEG. As the patients grew older, the seizures continued and evolved to Lennox-Gastaut syndrome (Fig. S1). Severe developmental delay was already evident at the first few months of life, predating the seizure disorder. No developmental regression was noted; however, the ultimate developmental outcome was poor and fell within the range of profound mental retardation. Growth parameters and general health were satisfactory, and there were no clinical or biochemical indications of other systems involvement. Numerous investigations including cerebrospinal fluid (CSF) profile, metabolic studies, skin biopsy for electron microscopy, and brain magnetic resonance imaging (MRI), were all normal. As an illustrative case the data are given for Patient 4696 (Data S1).
Table 1. Patients’ data and clinical course
Onset of seizures/diagnosis of West syndrome/onset of Lennox-Gastaut syndrome
Response to antiepileptic drugs
Last seizure type
F, female; M, male; y, effective; n, not effective.
Homozygosity mapping was performed in the samples of Patients 4695 and 4713, using the GeneChip Human Mapping 250K Nsp Array of Affymetrix (Affymetrix, Santa Clara, CA, U.S.A.) as previously described (Edvardson et al., 2007).
Consent was granted by the parents and the study was approved by the Hadassah ethical review committee.
Whole exome sequencing was performed in DNA sample of one patient, 4696, using the TargetSeq Exome enrichment kit on SOLiD 5500XL system (Life Technologies, Carlsbad, CA, U.S.A.) as 50-bp fragment reads plus Exact Call Chemistry (ECC). The sequence reads were processed with LifeScope Genomic Analysis Software (version 2.1; Life Technologies software products). These methods are detailed in the Data S1.
Functional studies of the mutation
In order to gain insight into the functional consequences of the mutation, recombinant soluble forms of the mutant and wild-type enzyme were generated and expressed in Chinese hamster ovary (CHO) cells. Protein secretion was monitored by enzyme-linked immunosorbent assay (ELISA), and the enzymatic activity was measured in a radioactive incorporation assay as described (Hu et al., 2011). These methods are detailed in Data S1.
The composition of the family, with healthy consanguineous parents, healthy sibs, and patients of both sexes suggested that WS in this family is caused by a founder mutation that is transmitted in an autosomal recessive manner. Using DNA single nucleotide polymorphism (SNP) chips in the samples of two affected cousins, a single common homozygous region was identified, spanning chr1: 30098510–45167143 (numbering according to HG19). We genotyped all family members for short tandem repeats (STR) and SNP markers across the region but could not narrow it down any further. Within the identified 15.1 Mb region there were 196 protein coding genes that contained 1,971 exons. Because no immediate candidate could be discerned, we opted for whole exome analysis of one of the patients (4696). This analysis, which had an average depth of coverage within targets of 49.59×, resulted in the identification of 33,728 SNP and 2,534 Indel. Of those, 180 homozygous SNP and five homozygous indel variants were found within the 15.1 Mb linked region. However, only five SNP were covered >×7 and were not present in dbSNP129, and only one was nonsynonymous and predicted by the pathogenicity prediction software Mutation Taster (Schwarz et al., 2010), to be disease causing.
This mutation resided on chr1: 44386520, c. 958 G>C, in exon 12 of the ST3GAL3 gene (gi1705561) corresponding to p. Ala320Pro (numbering according to the predominantly expressed isoform B1 [Q11203.1]). The sequence of exon 12 of ST3GAL3 was determined by Sanger sequencing in all family members. This analysis revealed that all the patients but none of the healthy sibs were homozygous for the mutation (Fig. 1A–D). The mutation was not carried by 123 anonymous controls of the same ethnic origin, and it was not found in the presently 5,379 exome analyses from healthy individuals available through the Exome Variant Server, NHLBI Exome Sequencing Project (ESP), Seattle, WA (http://evs.gs.washington.edu/EVS/) [04, 2012].
The gene ST3GAL3 encodes the β-galactoside-α2,3-sialyltransferase-III (ST3Gal-III), a Golgi resident type II transmembrane protein (Harduin-Lepers et al., 2001). Hallmarks of all vertebrate sialyltransferases are four highly conserved sialyl-motifs that form parts of the active site (Datta & Paulson, 1995; Jeanneau et al., 2004). The identified mutation p.Ala320Pro affects the sialyl-motif S, crucially involved in binding of both donor and acceptor substrate (Audry et al., 2011).
Analysis of the functional consequences of the Ala320Pro mutation revealed that the secretion of the mutant enzyme was reduced to 25% of the wild-type enzyme (Fig. 2B) and the isolated mutant protein showed no detectable activity (Fig. 2A). These data demonstrate that the Ala320Pro mutation drastically impacts or, more likely, completely abolishes ST3Gal-III function.
Four patients with WS and severe psychomotor retardation were investigated by linkage analysis followed by exome sequencing. This analysis revealed a homozygous missense mutation in the ST3GAL3 gene. ST3GAL3 encodes a member of the family of sialyltransferases (STs) which in humans comprises 20 enzymes (Harduin-Lepers et al., 2001). All STs use cytidinmonophophate (CMP) activated-sialic acid as donor substrate; however, they vary with respect to the recognized acceptor and the type of glycosidic linkage formed (Harduin-Lepers et al., 2001; Audry et al., 2011). The six STs of the ST3Gal subfamily transfer sialic acids in α2,3-linkage to galactose, and four of them, including ST3Gal-III, are involved in the synthesis of sialyl-Lewis (sLea and sLex) epitopes (Kono et al., 1997; Ellies et al., 2002). Although well known for their role in the leukocyte adhesion cascade, sialyl-Lewis epitopes have only recently been implicated in brain development through the identification of homozygous ST3GAL3 mutations in patients with nonsyndromic autosomal recessive intellectual disability (Hu et al., 2011). These mutations were localized outside the highly conserved sialyl-motifs and caused no clinical symptoms apart from intellectual disability. In contrast, the Ala320Pro mutation reported in the current study localizes in the functionally essential sialyl-motif S and fully destroys the enzymatic activity. In the mouse, ST3Gal-II and ST3Gal-III are responsible for nearly all of the terminal sialylation of brain gangliosides (Sturgill et al., 2012).
Therefore, the ST3GAL3 gene now joins the list of WS-associated genes by an as yet unknown mechanism. The present report underscores the usefulness of linkage analysis followed by exome sequencing in disease causing gene identification in small, consanguineous families.
We are grateful to Noa Cohen, Rachel Dahan, Michal Meimon, and Maike Hartmann for excellent technical assistance. This work was supported by the Hannover Biomedical Research School (HBRS) and by the Joint Research fund of the Hebrew university and Hadassah Medical Organization.
None of the authors has any conflict of interest to disclose. We confirm that we have read the Journal’s position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.