Address correspondence to Ulrich Stephani, Department of Neuropediatrics, University Medical Center Schleswig Holstein, Campus Kiel, Schwanenweg 20, 24105 Kiel, Germany. E-mail: email@example.com
Seizure disorders of the rolandic region comprise a spectrum of different epilepsy syndromes ranging from benign rolandic epilepsy to more severe seizure disorders including atypical benign partial epilepsy/pseudo-Lennox syndrome, electrical status epilepticus during sleep, and Landau-Kleffner syndrome. Centrotemporal spikes are the unifying electroencephalographic hallmark of these benign focal epilepsies, indicating a pathophysiologic relationship between the various epilepsies arising from the rolandic region. The etiology of these epilepsies is elusive, but a genetic component is assumed given the heritability of the characteristic electrographic trait. Herein we report on three patients with intellectual disability, various dysmorphic features, and epilepsies involving the rolandic region, carrying previously undescribed deletions in 16p13. The only gene located in the critical region shared by all three patients is GRIN2A coding for the alpha-2 subunit of the neuronal N-methyl-d-aspartate (NMDA) receptor.
The idiopathic focal epilepsies of childhood comprise a broad spectrum of phenotypes ranging from benign rolandic epilepsy (BRE) to more severe seizure disorders including atypical benign partial epilepsy/pseudo-Lennox syndrome (PLS), electrical status epilepticus during sleep (ESES), and Landau-Kleffner syndrome (LKS).
Centrotemporal spikes (CTS), well-structured five-phasic sharp-slow wave complexes, are the characteristic electroencephalographic pattern of these conditions. A strong genetic component of this electroencephalography (EEG) trait has been shown in twin and family studies (Doose & Baier, 1989). Despite the strong genetic contribution to the EEG trait, the etiology of epilepsies with centrotemporal spikes is largely unknown. Linkage has been described to the chromosomal region 15q14 (Neubauer et al., 1998), and rare variants in KCNQ2 and KCNQ3 can be identified in a small subset of patients (Neubauer et al., 2008). Mutations in SRPX2 have been identified in several disorders involving the rolandic region, including patients with BRE-like features (Roll et al., 2006). Together, however, these variants and mutations account for only a small part of the genetic risk for the rolandic epilepsies. Recently, an association with common variants within the ELP4 gene has been described, but not yet confirmed (Strug et al., 2009). Genetic findings, hinting at possible mechanisms of epileptogenesis in these patients, are, therefore, important to gain insight into the biology of these conditions.
The importance of structural genomic variants in seizure disorders is increasingly recognized. In several neuropsychiatric disorders including autism and intellectual disability, possibly pathogenic structural genomic variations are found in up to 10% of all patients, highlighting the importance of structural genomic variants (Sebat et al., 2007).
Herein we describe the case of a 6-year-old patient with various dysmorphic features, severe intellectual disability, and PLS carrying a deletion on 16p13.2p13.13 encompassing GRIN2A, coding for the alpha-2 subunit of the neuronal N-methyl-d-aspartate (NMDA) receptor. Overlapping deletions were also detected in a patient with unclassified epilepsy, severe intellectual disability, and an ESES-like EEG trait, as well as in a patient with developmental delay, several dysmorphic features, and a seizure disorder reminiscent of BRE.
The patient is the second child of two healthy and unrelated parents. Family history is unremarkable. The patient was born at 34 weeks of gestation with Apgar values of 3-7-7 and an umbilical artery pH of 7.34, as a product of a pregnancy complicated by oligohydramnios and intrauterine growth retardation. Global developmental delay was present from birth. When the child was 5 years of age, results of a neuropsychological assessment showed a level of a 1 year 1 month old. At the age of 6, expressive speech was still absent. On examination, mild facial dysmorphisms, brachydactyly, and clinodactyly were noted. Magnetic resonance imaging (MRI) and magnetic resonance spectroscopy (MRS) were unremarkable except for minor periventricular leukomalacia near the posterior horns, anatomically distant from the seizure focus and considered unrelated to the epilepsy phenotype. An extended metabolic screening including γ-aminobutyric acid (GABA) concentration in cerebrospinal fluid was within normal limits.
Atypical absences occurring at a frequency of 3–4 per day were first noted at the age of 2 years. Atonic seizures started by the age of 2½ with an initial frequency of 2–3 per day. Prolonged obtundation states for up to several days with eyelid and perioral myoclonias started at the age of 3½ years. At the age of 4 years, worsening of seizures, developmental slowing, and regression was documented. Generalized tonic–clonic seizures first occurred at the age of 4 years. EEG recordings were routinely obtained from the age of 12 months onward, showing bilateral independent centrotemporal spikes on an otherwise normal background. Starting at the age of 4 years, sleep recordings showed electrical status epilepticus in sleep.
Patient 2 was born in an uneventful pregnancy with a birth weight of 4000 g (P95). Length and head circumference were reported to be at the same percentile. Hypotonia and several dysmorphic features were observed at birth (Table 1). Mild to moderate intellectual disability was noted early on. At the age of 11 years, he was diagnosed with seizures reminiscent of rolandic epilepsy, for which he was temporarily treated with carbamazepine. The patient undergoes treatment for hypothyroidism.
Table 1. Phenotypic features of patients with deletions spanning GRIN2A
Patient 1 (age 7)
Patient 2 (age 20)
Patient 3 (age 12)
ESES, electrical status epilepticus in sleep; cm, centimeter; kg, kilogram; SD, standard deviation; CT, computed tomography; MRI, magnetic resonance image; MRS, magnetic resonance spectroscopy; chr, chromosome; FISH, fluorescence in situ hybridization.
Bilateral independent centrotemporal spikes, electrical status epilepticus in sleep (ESES)
Bilateral independent centrotemporal spikes reminiscent of ESES
Mild to moderate
Absent, mainly nonverbal communication
Delayed speech development
6 years 108 cm (−2 SD)
20 years 182.5 cm (+1 SD)
12 years 6 months 141 cm (−1.6 SD)
6 years 18.5 kg (−1.1 SD)
20 years 95.9 kg (+2 SD)
12 years 9 months 33.9 kg (−1.1 SD)
6 years 51.5 cm (−0.5 SD)
20 years 62.5 cm (+3.3 SD)
12 years 9 months 48 cm (−4 SD)
Cryptorchism, choanal stenosis
Club feet, cryptorchism
Walking since age 2
No walking, able to sit and roll over
Low-set ears, epicanthus, mild hypertelorism, widow′s peak, low-set backward rotated ears with overfolded helices and poorly formed anthelices brachydactyly, clinodactyly
Small and long face, mildly downslanting palpebral fissures, small narrow eyelids, broad nasal tip, long ears with large ear lobules, broad gums, broad neck, truncal obesity
Deep-set eyes, short nose with a broad nasal root, thin upper lip red., slight brachydactyly with tapering
Minor periventricular leukomalacia in the posterior horns MRI and MRS otherwise unremarkable
CT scan unremarkable
Mild diffuse reduction of brain volume on MRI
FISH clones deleted in the patients
RP11-544J16, RP11-418i22, RP11-107G6
Patient 3 is the first of four healthy children, reportedly born prematurely with meconium-stained amniotic fluid. Further information on birth measurements and perinatal history are not available. Developmental delay and several dysmorphic features were noted in the first year of life (Table 1). Epilepsy with polytopic myoclonic seizures and eyelid myoclonias was diagnosed at the age of five. Electroencephalographic recordings reveal bilateral independent centrotemporal spikes reminiscent of ESES.
Genome-wide screening for structural genomic variants was performed using the Human Genome CGH Microarray 44 K platform (Agilent Technologies, Santa Clara, CA, U.S.A.; Agilent 44k arrays in Patient 1 and Agilent 105k in Patient 3). For the analysis 1 μg of the patient’s DNA isolated from peripheral blood and 1 μg of reference DNA were hybridized using the manufacturer’s protocol with slight modifications. Statistically significant imbalances spanning a region of 10 sequential probes with a transformed Log2 ratio beyond 0.5 and mapping outside known copy number variations (CNVs) were considered aberrant. Mapped of array targets to the UCSC genome browser release March 2006 (http://genome.ucsc.edu/) revealed a 2.6 Mb deletion in patient 1 (7.9–10.5 Mb) and a 1.9 Mb deletion in patient 3 (9.3–11.2 Mb).
Screening for structural genomic variations in patient 2 was performed on DNA isolated from peripheral blood lymphocytes using the 32K BAC array platform with standard protocols, identifying a deletion in 16p13.2. This deletion was fine mapped by rehybridizing patient material to the 250K StyI SNP array from Affymetrix, identifying a ∼980 Kb deletion in 16p13.2 spanning the first 4–8 exons of GRIN2A. In all comparative genome hybridization (CGH) experiments the control consisted of pooled DNA of 10 males with a normal karyotype. An overview over the range of the various deletions is shown in Fig. 1. In all patients, deletions were confirmed by fluorescence in situ hybridization (FISH) analysis (details in Table 1). FISH analysis in parents of patients 1 and 2 revealed normal results, suggesting de novo deletions in both cases. Parental genotypes were not available for patient 3. No additional possibly pathogenic structural genomic variants were identified in patients 1–3 (list of identified CNVs available upon request, data on array CGH and FISH analysis are provided in the Supporting Information).
To assess the importance of 16p13.2p13.13 deletions in rolandic epilepsies, 63 patients with other epilepsies arising from the rolandic region (41 BRE, 12 PLS, 7 ESES, 3 LKS) were screened for deletions in GRIN2A using quantitative PCR (TaqMan, primers available on request). No deletions in GRIN2A were identified in any of the patients. Deletions of 16p13.2p13.13 involving exons of the GRIN2A gene are not listed in the Toronto Database of Genomic Variants (http://www.projects.tcag.ca/variation/), suggesting that the identified deletions have pathogenic significance.
In the present study we have identified three patients with a deletion in 16p13.2p13.13 spanning the GRIN2A gene. In addition to variable dysmorphic features and intellectual disability, all three patients have electroclinical features reminiscent of idiopathic focal epilepsies from the rolandic region.
Assuming that the disorder of the patients described in this manuscript represents a common syndrome, the smaller deletion in patient 2 narrows down the critical region for the epilepsy phenotype on 16p13.2p13.13 to the GRIN2A gene. GRIN2A codes for the alpha-2 subunit of the NMDA receptor and is implicated in glutamatergic transmission in the central nervous system (CNS). Glutamate as the major excitatory neurotransmitter in the CNS is critically involved in the generation of seizures, and dysregulation of glutamatergic neurotransmission has been shown in several animal models (Noebels, 2003). Given the overlapping deletions in the three patients described, we suggest that GRIN2A haploinsufficiency might contribute to the phenotype. However, positional effects on neighboring genes or long-range effects might also be causally related to the phenotype.
Despite dysmorphic features and intellectual disability, the epilepsy phenotype of patient 1 was compatible with PLS, including characteristic seizure types and the defining EEG trait. The larger deletion in patient 1 also included the ABAT gene. ABAT is a key enzyme in GABA metabolism, and ABAT mutations have been described in severe seizure disorders (Jaeken et al., 1984; Medina-Kauwe et al., 1999), making this a second candidate gene for the epilepsy phenotype in patient 1. However, GABA levels in the cerebrospinal fluid were normal, and molecular analysis failed to identify a pathogenic mutation of the second ABAT allele. In addition, deletions in patients 2 and 3 did not include the ABAT gene. Deletions of other additional genes in patients 2 and 3 might be responsible for the more severe phenotype and additional features.
In summary, we suggest that 16p13.2p13.13 deletions constitute a novel microdeletion syndrome with variable dysmorphic features and intellectual disability. Seizure disorders within the rolandic spectrum might represent a unifying feature of this syndrome.
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. None of the authors has any conflict of interest to disclose.