Spectrum of phenotypes in female patients with epilepsy due to protocadherin 19 mutations


Address correspondence to Nicola Specchio, Division of Neurology, Bambino Gesù Children’s Hospital – IRCCS, P.zza S. Onofrio 4, 00165 Rome, Italy. E-mail: nicola.specchio@opbg.net


Purpose: To describe clinical and neuropsychological features of six consecutive sporadic girls with protocadherin 19 (PCDH19) mutations.

Methods: Following recent descriptions of PCDH19 mutation in girls with epilepsy, we sequenced this gene in patients with infantile or early childhood seizures onset, either focal or generalized, without an obvious etiology.

Key Findings: Mean age at the time of the study was 13.5 ± 11 years. Mean age at seizure onset was 15.5 ± 11 months (range 9–38). All patients experienced clusters of either focal or generalized seizures, precipitated during febrile illness in five patients. Attacks were very frequent at onset, but they became less numerous during follow-up. Ictal electroencephalography (EEG) showed temporal lobe involvement in five patients. Periictal EEG showed focal or multifocal epileptiform and slow abnormalities. Cognitive impairment became obvious after seizure onset in three patients and was associated with autistic features in two. Genetic analysis revealed five new and one known de novo PCDH19 mutation that were missense in four and frameshift in two. Variants are clustered in the large exon 1, corresponding to the extracellular domain of the PCDH19 protein.

Significance: Our findings emphasize that de novo PCDH19 mutations are associated with infantile or early childhood onset of febrile or afebrile seizures often occurring in clusters. Cognitive impairment is not constantly present and autistic features are observed in some patients. Most patients have a “stormy” seizure onset, often related to fever; however, seizure severity does not clearly correlate with the degree of cognitive deficit. PCDH19 is likely a major epilepsy gene; phenotypes associated with mutations of this gene range from epileptic encephalopathies to mild epilepsy, yet large series of patients will be necessary to fully delineate phenotypic spectrum.

Mutations of PCDH19 were originally associated to epilepsy, with mental retardation limited to females (EFMR) (MIM 300088) (Dibbens et al., 2008). EFMR is a condition characterized by seizure onset in infancy or early childhood (6–36 months) and cognitive impairment. Seizures are predominantly generalized, including tonic–clonic, absence, myoclonic, tonic, and atonic seizures. Focal seizures have also been reported (Marini et al., 2010). The disorder is X-linked with an unusual expression pattern and with the phenotype being restricted to females; carrier males are apparently unaffected, with normal cognitive functions (Juberg & Hellman, 1971; Fabisiak & Erickson, 1990; Ryan et al., 1997; Scheffer et al., 2008). The spectrum of phenotypes associated with PCDH19, de novo or familial, mutations has been extended to include female patients with early onset epileptic encephalopathies resembling Dravet syndrome (DS) or focal epilepsy with or without mental retardation (Depienne et al., 2009; Hynes et al., 2010; Marini et al., 2010).

We describe six new sporadic girls with PCDH19 mutations and attempt phenotype–genotype correlations. Our findings confirm that PCDH19 causes infantile or early childhood onset epilepsy in females patients. Most patients have a “stormy” seizure onset with clusters of focal or generalized seizures, often triggered by febrile episodes. Cognitive impairment and autistic features are often present but are by no means constant. A better definition of the spectrum of phenotypes associated with PCDH19 mutations will help clinicians to obtain an early molecular diagnosis that is relevant for the patient’s management, to help to avoid useless invasive and expensive diagnostic testing and to improve prognosis and genetic counseling.


We describe clinical, neurophysiologic, neuropsychological, and neuroimaging data of six female patients with PCDH19 mutations. We screened 75 consecutive girls with epilepsy. Following our previous study (Marini et al., 2010) in which we identified familial and de novo PCDH19 mutations in girls with early onset cryptogenic focal epilepsy, extending the spectrum of phenotypes beyond DS and EFMR, we screened a series of female patients with a wide spectrum of epilepsy phenotypes including either focal or generalized, febrile or afebrile seizures, with normal MRI, and with or without mental retardation of unknown origin. Most patients had an epilepsy age of onset within the first few years of life, yet a few patients with a childhood onset were included.

We undertook this study without any ambition of providing a percentage of the contribution of PCDH19 mutations to the etiology of infantile epilepsy; we limited our purpose to increasing the number of descriptions and, in so doing, moving forward in the understanding of the phenotypic spectrum. In those patients in whom we detected pathogenic mutations, we reviewed medical records, family history, seizure semiology, susceptibility to fever, frequency of seizure and duration (isolated, cluster, status epilepticus), treatment of clusters and/or status epilepticus, interictal and ictal electroencephalography (EEG), brain imaging, treatment during the follow-up, and cognitive and behavioral assessment. Epileptic seizures were classified according to the International League Against Epilepsy criteria (Blume et al., 2001; Engel, 2001). We reviewed all EEG studies performed during follow-up. The Scientific Directorate of our Hospital according to the local regulations approved the study.

Mutation analysis

Genomic DNA was extracted from peripheral blood leukocytes of patients and available parents. Genomic DNA was extracted by QUIAMP DNA Blood mini kit, according to the manufacturer’s protocol (Quiagen, Hilden, Germany). The six exons covering the coding regions of PCDH19 (Entrez Gene, GeneID: 57526, accession number: EF676096.1) and their respective intron–exon boundaries were amplified by polymerase chain reaction (PCR) and cycle sequenced using BigDye 3.1 chemistry on an ABI 3130xl (Applied Biosystems, Foster City, CA, U.S.A.). Due to the large size, PCDH19 exon 1 was amplified and cycle sequenced using five internal primers. The new identified PCDH19 mutations were checked in a control population of 150 ethnically matched controls, and segregation analysis was performed on all available members of a pedigree.

X-inactivation studies

Five hundred nanograms of genomic DNA was digested overnight at 37°C in a volume of 50 μl containing 1× digestion buffer and 1 μl of the HpaII enzyme (10 U/μl) (Roche Diagnostics, Mannheim, Germany). After digestion, the enzyme was inactivated at 65°C for 20 min, and PCR (30 cycles) was performed on both digested and undigested DNAs (200 ng) with primers specific for the androgen receptor gene. The forward primer was marked at the 5′ with FAM. The fluorescent PCR was analyzed and sized using the GeneScan – 500 LIZ Size Standard (Applied Biosystems) as internal reference on the 3130XL ABI Prism DNA sequencer (Applied Biosystems). X-inactivation was classified as random (ratio 50:50–80:20) or skewed (ratio >80:20).


Demographic and clinical characteristics of the six female patients harboring PCDH19 mutations are summarized in Table 1.

Table 1.   Summary of the clinical and genetic data of the six female patients with PCDH19 mutations
Patient no.Age at study/sexAge at onsetSeizure type at onsetSeizure frequency at onsetSEFeverInterictal EEGIctal EEGTreatment (follow-up)Seizure type (follow-up)Cognitive and behaviorPCDH19 mutations (inheritance)
  1. y, years; m, months; F, female; SE, status epilepticus; GTC, generalized tonic–clonic; SG, secondary generalization; abn, abnormalities; SW, slow waves; BA, background activity; PS, polyspikes; ShW, shrap waves; R, right; L, left; F, frontal; T, temporal; C, central; P, parietal; O, occipital; PPR, photoparoxysmal response; bil, bilateral; PB, phenobarbital; CBZ, carbamazepine; GVG, gamma-vinyl-GABA; CLB, clobazam; CZP, clonazepam; PHT, phenytoin; DZP, diazepam; VPA, valproate; ETS, ethosuximide; LTG, lamotrigine; TPM, topiramate; LEV, levetiracetam; ACZ, acetazolamide; IQ, intelligent quotient; GQ, general quotient; WPPSI, Wechsler Preschool and Primary Scale of Intelligence.

117 y 6 m/F9 mGTC, focalClusterYesNoFocal epilept abn; R T-O SWDiffuse and focal onsetPB, CBZ, GVG, CLB, CZP, VPA, ETS, PHT, LTG, TPM, LEVCluster of GTC or focal (3–4/y)IQ 64 (8 y)
Behavioral disturbances
(de novo)
21 y 11 m/F10 mFocal with SGClusterNoYesClusters: slow BA
sleep: rare diffuse SW
Vertex fast PSPB, LEV, CLBCluster of focal (several/y)GQ 89.2 (1 y 7 m) (Griffith)c.1537G>C
(de novo)
335 y 5 m/F10 mGTC, tonicClusterNoYesBil FC theta; rare ShWDiffuse dischargeVPA, PB, CZP, CLB, GVG, ACZ, CBZ, DZPFocal with SG
Atypical absences
(from 13 y, 3–4/y)
IQ 48 (15 y)
Autistic features
(de novo)
47 y 3 m/F1 yFocal with SGClusterNoYesNormalL T-P fast PSGVG, PB, VPA, TPM, LEVCluster (several/y)GQ 89 (3 y 3 m) (Griffith)
IQ 86 (6 y) (WPPSI)
(de novo)
511 y 10 m/F14 mGTCClusterNoYesSlow BA, bil FT, OT focal epilept abn, PPRDiffuse and focal onsetPB, VPA CNP, TPM, CLBFocal, tonic or GTC (monthly/yearly)IQ 50 (7 y)
IQ 35 (9 y 6 m)
Autistic features
c.1300_1301delCA p.Gln434GlufsX11
(de novo)
67 y 6 m/F3 y 2 mTonic, focalClusterNoYesBil F thetaDiffuse and focalVPA, CBZ, LEV, PyridoxineCluster or isolated focal (1–2/y)GQ 98 (3 y 6 m) (Griffith)
IQ 101 (5 y 9 m) (WPPSI)
c.958dup(G p.Asp320GlyfsX22)
(de novo)

The mean age at the time of the study was 13.5 ± 11 years (range 1.11–35.5 years). Three patients had a family history for febrile seizures (FS) and/or epilepsy (Table 1: Patients 3, 5, and 6). Mean age at seizure onset was 15.5 ± 11 months (range 9–38). Five patients had seizures precipitated by fever (Table 1: Patients 2–5). All patients had either focal or generalized seizures. In all patients, at onset, seizures appeared in clusters, whereas during follow-up, in all but one, attacks continued to manifest as clusters having a variable frequency (monthly to yearly). Three patients (Table 1: Patients 3, 5, and 6) also experienced isolated seizures. Attacks tended to decrease in frequency during follow-up, and two patients (Table 1: Patients 1 and 3) had been seizure-free for 6 and 5 years. All patients had at least one ictal EEG recorded: Ictal discharge involved the temporal regions in five (temporooccipital, temporoparietal, centrotemporal, and posterior temporal), and the vertex in one. Four patients (Table 1: Patients 1, 3, 5, and 6) had also at least one generalized seizure documented (Fig. 1A,B). During the cluster of seizures interictal EEGs showed: slow background activity in four patients (Table 1: Patients 1, 2, 5, and 6); focal or multifocal epileptiform abnormalities in two (Table 1: Patients 1 and 5); and intermittent, multifocal, theta activity in four (Table 1: Patients 1, 2, 3, and 6). A photoparoxysmal response was detected in one patient at the age of 7 years (Table 1: Patient 5) (Fig. 2A,B).

Figure 1.

Ictal EEG showing a tonic postural (A) and focal (B) seizure. (A) Patient 1: a diffuse low voltage fast activity followed by polyspikes involving all traces, and by a generalized spikes-and-wave discharge. (B) Patient 6: A focal seizure recorded during a cluster. Seizure was recorded during a long-term monitoring during phase II of non–rapid eye movement (NREM) sleep. A theta activity over right frontotemporal region is evident. The patient opens eyes with fear expression and asks for help; this is followed by left head orientation. After 12 s the ictal discharge involves also the right central and frontal areas. The seizure lasts almost 40 s. At the end, chewing automatisms are evident (see muscle artifact over temporal and central traces).

Figure 2.

Photoparoxysmal epileptiform response in Patient 5. Since the age of 7 years, intermittent photic stimulation elicited a photoparoxysmal response not associated with clinical events. Frequencies of stimulations (A) 18 Hz and (B) 20 Hz.

Psychomotor development, prior to the onset of epilepsy, was reported as normal in all. Three patients manifested various degrees of cognitive impairment after seizure onset (Table 1: Patients 1, 3, and 5), which was associated with autistic features in two (Table 1: Patients 3 and 5) and behavioral disturbances in one (Patient 1). Brain MRI was normal in all. Epilepsy was refractory to antiepileptic drugs (AEDs) at onset in all patients, and clusters of seizures recurred frequently despite treatment. The frequency of clusters of seizures reduced over time without a clear correlation with AED adjustment. Indeed, two patients (Table 1: Patients 1 and 3) had had 5–6 years of seizure-free periods without modifying the treatment.

Genetic analysis revealed five new, de novo PCDH19 mutations (p.Leu433Pro, p.Gly513Arg, p.Pro236Ser, p.Gln434GlufsX11, and p.Asp320GlyfsX22) and one known de novo variant (p.Asn340Ser). All variants were clustered in the large exon 1, corresponding to the extracellular cadherin domain of the protocadherin 19 protein (Fig. 3). Four patients had a variable X-inactivation within the random pattern. For two patients, X-inactivation studies could not be performed (Table 1: Patients 1 and 6). Details for each patient are available as Supporting Information.

Figure 3.

Graphic representation of the PCDH19 protein reporting all the mutations reported in literature (bottom) and mutations identified in this study (top). SP, signal peptide; EC, extracellular cadherin domain; TM, transmembrane; CP, cytoplasmic domain.


PCDH19 is emerging as a major gene in female patients with infantile onset epilepsy with either familial clustering or de novo mutations. Clinical features of patients so far described include infantile onset of multiple seizure types, especially focal with possible evolution into status epilepticus. Seizures are often triggered by febrile illnesses; patients have either normal cognitive functions or mental retardation, with or without autistic features (Scheffer et al., 2008; Depienne et al., 2009; Marini et al., 2010). Clinical similarities with DS have been suggested in some patients, (Depienne et al., 2009; Marini et al., 2010), yet females with normal cognitive functions and focal epilepsy might also carry PCHD19 mutations (Marini et al., 2010).

We identified pathogenic de novo PCDH19 mutations in six girls with clusters of focal and generalized seizures, followed in three girls by mental retardation, autistic features, and behavioral disturbances. Mean age at seizure onset was 15.5 months, ranging from 9 months to 3 years and 2 months, extending the spectrum of PCDH19-associated-epilepsy from infantile to early childhood onset. All patients had multiple seizure types including: tonic (symmetrical or asymmetrical) tonic–clonic, and focal seizures, often occurring in clusters and triggered by fever in five.

As previously reported (Marini et al., 2010), the occurrence of seizures in cluster was a constant feature also in our patients. At onset, clusters are more frequent, resistant to AED, prolonged, sometimes evolving into status epilepticus. Such a “stormy onset,” notably in the presence of fever, might induce clinicians to consider encephalitis. Over time clusters tended to decrease in frequency but patients are still at risk of experiencing status epilepticus, even during childhood or adolescence.

The six patients here described had a long follow-up, ranging from 1–34 years. Clinical outcome shows that this condition might be considered as “semiprogressive” meaning that in the first few years of life seizure frequency increases, cognitive impairment worsens, and autistic or behavioral disturbances may appear. However, over time seizures become less frequent and cognitive level plateaus, leading to an overall outcome that is less severe than in other epileptic encephalopathies the onset of which is of comparable severity. Indeed the oldest patient of our series was 35 years of age and is currently experiencing only rare, isolated atypical absences and focal seizures, with steady cognitive abilities and autistic features. At onset, clusters of seizures were refractory to AEDs and required intravenous midazolam and intensive care admissions. In between clusters a combination of two or more AEDs was used; patients were seizure free for months or years, yet treatment did not prevent the recurrence of other clusters of seizures, and there were no data of a better efficacy of one AED compared to others. Clinical evidence and long-term follow-up suggest that seizure improvement is related to growth and reduced episodes of febrile illnesses rather than to treatment.

Ictal EEGs showed a focal origin of seizures with involvement of the temporal regions in five patients; seizures with a bilateral and diffuse onset were also recorded. Periictal EEGs showed slow background activity and focal epileptiform discharges, with either left or right predominance. The long-term follow-up in the two oldest patients (16 and 35 years of age) revealed that interictal EEG, outside clusters of seizures, became normal. In one of our patients (Table 1: Patient 5) a photoparoxysmal response was observed. Overall, EEG features did not appear sufficiently specific to guide toward PCDH19 mutation screening.

Early psychomotor development was seemingly normal in all patients. Following seizure onset, three patients appeared to have mild to moderate cognitive impairment; two of them also had autistic features. Our findings, therefore, confirm that most patients with PCDH19 mutations exhibit variable degrees of cognitive deficits, which are detected following seizure onset (Marini et al., 2010). Of course, mild cognitive impairment, predating seizure onset, would easily be overlooked. Our series also confirms that some patients maintain normal cognitive skills despite frequent clusters of seizures.

None of the patients with PCDH19 mutations exhibited a Dravet-like phenotype, but this finding is obviously dependent on the population studied. Seizure types, EEG, and cognitive features were not suggestive of EFMR either. Epilepsy phenotype of EFMR, as described by Scheffer et al. (2008), is that of infantile onset of predominantly generalized seizures, with variable outcome yet overall not very severe and with epilepsy offset around 12 years of age in a number of patients. Family history and pedigree inspection was also not indicative of EFMR.

Mutations were missense (4), and frameshift (2), five being new mutations. All reported variants are clustered in the large exon 1 of the gene, corresponding to the extracellular cadherin domain of the PCDH19 protein. Extracellular domains of protocadherins are crucial for normal function (Sano et al., 1993). Based on alignment score, all the variants described affect highly conserved amino acids in the protein (Fig. 4) and probably cause loss-of-function of the mutated allele. The p.Asn340Ser and p.Asp320GlyfsX22 mutations affect residues belonging to conserved calcium binding motifs (Patel et al., 2006). In particular, amino acid 320 seems to play a crucial role in the calcium domain, so these substitutions can either affect the calcium binding activity or alter the PCDH19 extracellular structure, other than causing a premature protein termination. Residues p.Leu433Pro, p.Gln434GlufsX11, p.Gly513Arg, and p.Pro236Ser are strongly conserved among δ1 and δ2 PCDHs suggesting a critical role in protein function. These mutations might be predictive of a more severe course than others: Indeed the three patients with associated mental retardation carried such mutations (see Table 1).

Figure 4.

Multiple sequence alignments of human protocadherins and orthologs of PCDH19 from other species were generated by Jalview editor (http://www.jalview.org). Residues were colored according to their physicochemical properties (Zappo color scheme). Black arrows indicate residues involved in missense substitutions.

The increasing reports of female patients carrying PCDH19 mutations will allow better phenotype–genotype correlations. The phenotypic spectrum is likely wider than currently known, and more patients need to be screened before the whole picture is clarified. Our findings emphasize that female patients with epilepsy beginning in infancy or early childhood, with clusters of febrile or afebrile, focal, or generalized seizures, with or without cognitive impairment or autistic features, are potential mutation carriers. PCDH19 mutations appear to cause a “stormy” seizure onset, often related to febrile episodes. Clusters of seizures tend to recur frequently during the first years of life; still seizure severity does not clearly correlate with the development and degree of cognitive deficits. Other genetics factors might come into play to “rescue” the phenotype into mild epilepsy in some patients. A molecular diagnosis is important for patient’s management, prognostic predictions, and genetic counseling.


The authors have no conflicts of interest to declare. 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.