Address correspondence to Toshiyuki Yamamoto, Tokyo Women’s Medical University Institute for Integrated Medical Sciences, 8-1 Kawada-cho, Shinjuku-ward, Tokyo 162-8666, Japan. E-mail: email@example.com
Purpose: Genetic mutations of the cyclin-dependent kinase-like 5 gene (CDKL5) have been reported in patients with epileptic encephalopathy, which is characterized by intractable seizures and severe-to-profound developmental delay. We investigated the clinical relevance of CDKL5 alterations in both genders.
Methods: A total of 125 patients with epileptic encephalopathy were examined for genomic copy number aberrations, and 119 patients with no such aberrations were further examined for CDKL5 mutations. Five patients with Rett syndrome, who did not show methyl CpG-binding protein 2 gene (MECP2) mutations, were also examined for CDKL5 mutations.
Key Findings: One male and three female patients showed submicroscopic deletions including CDKL5, and two male and six female patients showed CDKL5 nucleotide alterations. Development of early onset seizure was a characteristic clinical feature for the patients with CDKL5 alterations in both genders despite polymorphous seizure types, including myoclonic seizures, tonic seizures, and spasms. Severe developmental delays and mild frontal lobe atrophies revealed by brain magnetic resonance imaging (MRI) were observed in almost all patients, and there was no gender difference in phenotypic features.
Significance: We observed that 5% of the male patients and 14% of the female patients with epileptic encephalopathy had CDKL5 alterations. These findings indicate that alterations in CDKL5 are associated with early epileptic encephalopathy in both female and male patients.
Epileptic encephalopathies are a group of conditions in which neurologic deterioration results mainly from epileptic activity. The clinical and electroencephalography (EEG) characteristics depend on the age of onset and may change over time (Zupanc, 2009). An underlying genetic background has been suggested in patients with epileptic encephalopathy (Nabbout & Dulac, 2008). An X-linked gene coding for cyclin-dependent kinase-like 5 gene (CDKL5; MIM #300203) is one of the genes responsible for epileptic encephalopathy. Kalscheuer et al. (2003) identified de novo balanced X autosome translocations in two female patients with infantile spasms, in whom CDKL5 was disrupted. Since then, the phenotypic spectrum of CDLK5 abnormalities has expanded to include features resembling Rett syndrome (RTT; MIM #312750) with early onset seizures (Evans et al., 2005; Mari et al., 2005). Now, phenotypic features of CDLK5 abnormalities are widely recognized as early infantile epileptic encephalopathy-2 (EIEE-2; MIM #30062) and are characterized as severe epileptic encephalopathy associated with early onset and refractory seizures (Archer et al., 2006; Pintaudi et al., 2008).
Although the consequence of CDKL5 alterations has also been attributed to X-linked dominant infantile spasm syndrome-2 (ISSX2), mutations have been identified not only in female patients but also in some male patients with severe mental retardation and early onset intractable seizures (Elia et al., 2008; Fichou et al., 2009; Sartori et al., 2009). Therefore, we performed a comprehensive analysis for CDKL5 in both female and male patients with epileptic encephalopathy.
After obtaining approval of the study protocol by the ethics committee of the institution and informed consent from the families of the patients, peripheral blood samples of 125 patients (59 male and 66 female) with epileptic encephalopathy of unknown etiology were collected, together with their clinical information, including neuroimaging findings. Epileptic encephalopathies are defined as disorders in which there is a temporal relationship between deterioration in cognitive, sensory, and motor function and epileptic activity, which includes frequent seizures and/or extremely frequent interictal paroxysmal activity (Nabbout & Dulac, 2003). Five female patients with RTT who did not show methyl CpG-binding protein 2 gene (MECP2) mutations (which are often associated with RTT) were also included in the cohort study for CDKL5 mutations.
The genomic copy numbers of the patients with epileptic encephalopathies were determined using the Human Genome CGH Microarray 105K (Agilent Technologies, Santa Clara, CA, U.S.A.) as described previously (Shimojima et al., 2010).
Validation of the genomic copy number aberrations
Fluorescent in situ hybridization (FISH) analysis was performed for the large chromosomal deletion by using bacterial artificial chromosome (BAC) clones as probes, RP11-106N3 and CTD-2335C24 including CDKL5 as a target, and RP11-1051J20 as a marker (Fig. 1, Table S1). The deletion identified in Patient 1 was too small to be detected by a BAC clone; therefore, multiplex polymerase chain reaction (PCR) analysis was used for validation. Two DNA fragments, exon 1B (421 bp) and exon 2 (350 bp) of CDKL5, were amplified in the same PCR reaction tube, separated by agarose gel electrophoresis, and visualized by ethidium bromide staining.
Cohort study for CDKL5
Samples from 119 patients (58 male and 61 female) that showed no genomic copy number aberrations at the first screening by microarray-based comparative genomic hybridization (aCGH) in this study were included in the second cohort. Five samples obtained from female patients with RTT who did not show MECP2 mutations were also included. The genomic sequences of all 23 exons of CDKL5 were analyzed by the standard PCR direct-sequencing method using primers listed in Table S2. A recently identified exon 16B, which if included in the mature mRNA produces as a new CDKL5 isoform, was also analyzed in this study (Fichou et al., 2010). When nucleotide changes were identified in samples for which parental samples were available, trio analyses were performed to test whether the mutation was de novo or familial. DNA samples collected from 100 healthy Japanese volunteers (50 male and 50 female) comprised the control cohort.
Genomic copy number aberrations
In Patient 1, an aberration was identified at Xp22.13, indicating a nullisomy of this region (Fig. 1, Table S3). This region corresponds to exon 1 of CDKL5. Subsequent multiplex PCR analysis using two sets of primers for exon 1B and exon 2 of CDKL5 showed no band for exon 1B (Fig. 2A), thereby confirming the nullisomy of this region. Both parents of Patient 1 declined trio analysis.
aCGH analysis identified chromosomal aberrations in the CDKL5 region in three female patients (Fig. 1, Table S3). Because male reference DNA was used in this study, genomic copy numbers of the normal female X chromosome regions showed log2 ratio of +1. Therefore, a log2 ratio of “0” indicates the same genomic copy numbers with the male reference sample, indicating a partial monosomy of this region in these patients. For Patients 2 and 3, identified aberrations were confirmed by FISH by detecting only one signal with RP11-106N3 and CTD-2335C24, respectively, indicating deletions in this region (Fig. 2B,C). For Patient 4, one of the targeted signals of CTD-2335C24 was weaker than the other, indicating a partial deletion of the targeted region (Fig. 2D). For Patients 2 and 3, the deletion region involved four genes: CDKL5; X-linked juvenile retinoschisis protein gene (RS1), which is responsible for X-linked juvenile retinoschisis (MIM #312700); protein phosphatase with EF hand calcium-binding gene (PPEF1); and phosphorylase kinase alpha 2 gene (PHKA2), which is responsible for X-linked hepatic glycogen storage disease (MIM #300798). For Patient 3, the deleted region involved the latter half of CDKL5 after exon 4. Patient 4 also showed a partial CDKL5 deletion after exon 16, and RS1, which was encoded in the antisense direction. For Patients 2, 3, and 4, both parents were negative for these deletions, indicating de novo origin.
There were no other known pathogenic aberrations in these four patients. In the other two patients, genomic copy number aberrations in the region of the platelet-activating factor acetylhydrolase gene (PAFAH1B1), which is responsible for lissencephaly, were identified (Shimojima et al., 2010). The remaining 119 patients showed no genomic copy number aberrations and were included in the cohort study for CDKL5 mutations.
CDKL5 nucleotide alterations
In the 119 patients, eight pathogenic mutations were identified (including six novel and two recurrent mutations), which consisted of three nonsense mutations, three frameshift mutations, and two missense mutations (Fig. 3, Table 1). Aristaless-related homeobox gene (ARX; MIM #300382) was not found in any of the male patients. Five patients with RTT who did not show MECP2 mutations also did not show mutations in CDKL5. No control samples showed any of the nucleotide alterations identified in this study (Table 1).
Table 1. Summary of the clinical features and the identified CDKL5 mutations in the patients reported in this study
M, male; F, female; EE, epileptic encephalopathy; y, years; m, months; w, weeks; d, days; OFC, occipitofrontal circumference; NT, not tested; Sz, seizures; SPECT, single-photon emission computed tomography.
Although Patient 10 showed a nonsense mutation (p.S413X), an additional missense mutation (p.H467P) was also identified in exon 12. Neither alteration was found in parents, indicating de novo occurrence of both mutations. Because a similar missense mutation (p.H467R) was reported to be a nonpathogenic mutation, p.H467P is also expected to be a nonpathogenic mutation (Evans et al., 2005).
Brain magnetic resonance imaging (MRI) of the patients with CDKL5 alterations is shown in Fig. 4. Many patients showed frontal dominant cerebral atrophy. All clinical data including the findings of neuroimaging are summarized in Table 1. The ability to sit autonomously was the maximum gross motor development achieved by these patients, and none of the patients acquired speech ability, indicating severe developmental delay. Only the oldest patient (Patient 7; 4 years and 7 months old), who had a missense mutation, showed seizure control after 3 years of age; all the other patients had persistent seizures.
Using aCGH analyses, Erez et al. (2009) identified partial CDKL5 deletions in female patients with early onset intractable epilepsy. Mei et al. (2010) identified four patients who had total or partial deletions in CDKL5. However, those studies included only female patients. In comparison, the aim of our study was to identify candidate genetic causes of early epileptic encephalopathy, and thus we recruited patients of both genders. Genomic copy numbers of whole chromosomes were comprehensively analyzed and submicroscopic chromosomal abnormalities of the CDKL5 region were identified in both genders. The male patient (Patient 1) showed a partial deletion of CDKL5. Patients 2 and 3 showed large deletions in which the four neighboring genes, CDKL5, RS1, PPEF1, and PHKA2, were included. RS1 and PHKA2 are responsible for X-linked diseases, and the function of PPEF1 is unknown. The remaining Patient 4 showed partial deletions of CDKL5 and RS1. Therefore, phenotypic features of Patients 2, 3, and 4 suggest a causal role for CDKL5 deletions in early epileptic encephalopathy. Despite the gender difference and the deleted size differences, the clinical severities of the patients with CDKL5 deletions were similar between genders and similar to those of patients previously reported to have partial or total deletion of CDKL5 (Van Esch et al., 2007; Erez et al., 2009; Bahi-Buisson et al., 2010; Mei et al., 2010).
Previously, CDKL5 mutations were shown to affect mainly female patients, and their frequency has been estimated as approximately 9–28% in female patients with early onset seizures (Bahi-Buisson et al., 2008b; Nemos et al., 2009). However, those studies mainly included female patients. Elia et al. (2008) identified CDKL5 mutations in three male patients with early onset epileptic encephalopathy. Male patients with CDKL5 mutations or deletions have also been reported by others (Fichou et al., 2009; Sartori et al., 2009). In our study, initial identification of CDKL5 deletions in both male and female patients with early epileptic encephalopathy prompted us to analyze CDKL5 nucleotide sequences of both genders, and the results revealed nucleotide changes in two male patients and six female patients. We observed that the clinical severity of the disease did not differ between males and females. Therefore, male as well as female patients with early onset epileptic encephalopathy should be tested for CDKL5 mutations.
Because CDKL5 is located on Xp22.13, genetic traits of CDKL5 alterations have been considered to be X-linked dominant, just as MECP2 mutations are responsible for the majority of RTT cases, a neurologic disorder occurring almost exclusively in females. The rare male patients with MECP2 mutations showed severe mental retardation but no RTT phenotype (Gomot et al., 2003). In comparison, there are no phenotypic differences between male and female patients with CDKL5 mutations or deletions. Bahi-Buisson et al. (2008b) suggested that phenotypic heterogeneity does not correlate with the nature or the position of the mutations or with the pattern of X-chromosome inactivation. Indeed, no clear genotype–phenotype correlation between these factors has been established. Therefore, an important question is why clinical severity is the same between the genders. Based on previous reports, we know that the absence of CDKL5 protein is not lethal in males, and CDKL5 abnormalities result in severe neurodevelopmental delay and early onset epilepsy in both genders (Castren et al., 2011). In this study, the estimated frequencies of CDKL5 abnormalities in patients with epileptic encephalopathy were 5% in male and 14% in female patients. Therefore, the observed difference in the frequency of CDKL5 mutations between male and female patients may simply be a consequence of the fact that female patients have two X chromosomes.
Subjects in our study included five female patients with RTT who did not show MECP2 mutations. However, these female patients did not carry a CDKL5 mutation. Some researchers have found no CDKL5 mutations in patients with RTT (Huppke et al., 2005; Li et al., 2007). Previously, CDKL5 mutations were analyzed in patients with both classic and atypical variants of RTT. However, mutations were identified only in patients with seizure onset before 6 months of age (Evans et al., 2005; Scala et al., 2005; Artuso et al., 2010). In another study, all patients with CDKL5 mutations showed early onset seizures that began before 6 months of age (Erez et al., 2009). These findings suggest that development of early onset seizures is an essential clinical feature in patients with CDKL5 mutations. The onset of epileptic seizures in the first 6 months distinguishes patients with CDKL5 mutations from patients with typical RTT caused by MECP2 mutations (Castren et al., 2011).
All previously reported CDKL5 mutations were sporadic and were identified as de novo. Only a small numbers of mutations were recurrent (Castren et al., 2011). In this study, we observed eight CDKL5 mutations that included six novel and two recurrent mutations. The phenotypic features of the patients with recurrent mutations are similar to those described previously (Sartori et al., 2009; Artuso et al., 2010).
Consistent with the findings of previous studies, we observed polymorphous seizures (i.e., myoclonic seizures, tonic seizures, and spasms) in our study. The clinical course of seizure development was also identical to the proposed three stages reported by Bahi-Buisson et al. (2008a) [i.e., stage I, early onset epilepsy (onset 1–10 weeks); stage II, epileptic encephalopathy with infantile spasms and hypsarrhythmia; stage III, seizure-free in estimated 50% of patients at late infantile period] because our Patient 7 showed good seizure control after 3 years of age. Artuso et al. (2010) reported that patients with CDKL5 mutations showed no abnormalities on brain magnetic resonance imaging (MRI). However, our findings indicated mild frontal lobe atrophy in almost all patients. Therefore, this may be an additional clinical characteristic of patients with CDKL5 mutations.
We thank the patients’ parents for their gracious participation and support. J-S L was supported by a Research Fellowship from the Takeda Science Foundation in Japan. This research was partially supported by Research Grant from the Japan Epilepsy Research Foundation (T.Y.)
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.