Prospective study of POLG mutations presenting in children with intractable epilepsy: prevalence and clinical features

Abstract

Purpose: To assess the frequency and clinical features of childhood-onset intractable epilepsy caused by the most common mutations in the POLG gene, which encodes the catalytic subunit of mitochondrial DNA polymerase gamma.

Methods: Children presenting with nonsyndromic intractable epilepsy of unknown etiology but without documented liver dysfunction at presentation were eligible for this prospective, population-based study. Blood samples were analyzed for the three most common POLG mutations. If any of the three tested mutations were found, all the exons and the exon-intron boundaries of the POLG gene were sequenced. In addition, we retrospectively reviewed the notes of patients presenting with intractable epilepsy in which we had found POLG mutations. All available clinical data were collected by questionnaire and by reviewing the medical records.

Key findings: We analyzed 213 blood DNA samples from patients fulfilling the inclusion criteria of the prospective study. Among these, five patients (2.3%) were found with one of the three common POLG mutations as homozygous or compound heterozygous states. In addition, three patients were retrospectively identified. Seven of the eight patients had either raised cerebrospinal fluid (CSF) lactate (n = 3) or brain magnetic resonance imaging (MRI) changes (n = 4) at presentation with intractable epilepsy. Three patients later developed liver dysfunction, progressing to fatal liver failure in two without previous treatment with sodium valproate (VPA). Furthermore, it is worth mentioning that one patient presented first with an autism spectrum disorder before seizures emerged.

Significance: Mutations in POLG are an important cause of early and juvenile onset nonsyndromic intractable epilepsy with highly variable associated manifestations including autistic features. This study emphasizes that genetic testing for POLG mutations in patients with nonsyndromic intractable epilepsies is very important for clinical diagnostics, genetic counseling, and treatment decisions because of the increased risk for VPA-induced liver failure in patients with POLG mutations. We recommend POLG gene testing for patients with intractable seizures and at least one elevated CSF lactate or suggestive brain MRI changes (predominantly abnormal T2 -weighted thalamic signal) with or without status epilepticus, epilepsia partialis continua, or liver manifestations typical for Alpers disease, especially when the disease course is progressive.

Figure 1

Increased T₂ signal intensities in dentate nuclei and thalamic regions in brain magnetic resonance imaging (MRI) associated with POLG mutations. Axial T₂-weighted MR images from patient 4, who had p.W748S;E1143G + G1205E mutations. There is mildly increased signal intensity (SI) in the region of the dentate nuclei (A, arrow) and diffuse high SI in both thalami (B, arrow). Increased SI in the same areas was also seen on the fluid-attenuated inversion recovery (FLAIR) images but no abnormal findings were detected on the T₁-weighted images. No other lesions were seen in this patient and the findings were very similar between the scan at presentation (shown) and a scan done 3 months later.

Figure 2

Progressive changes in brain MRI with continued fitting associated with POLG mutations. The brain MRI at presentation with the first focal epileptic seizure of patient 7 with a homozygous p.W748S/p.W748S was normal (not shown). On admission 1 year later a clearly circumscribed region of abnormal high signal intensity (SI) was seen in the right posterior thalamus on T₂-weighted and FLAIR images (A, axial T₂-weighted image shown at the level of the basal ganglia) and abnormal low SI on T₁-weighted imaging (not shown). Lesions of similar SI were seen in two sites in the left cerebellar hemisphere (B, axial FLAIR image). Two months later, bilateral parietal-occipital lesions involving the white matter and cortex (right much more than left) were seen (C, axial T₂-weighted image) with improvement in the previous abnormalities without atrophy, although the abnormal SI in the left thalamus remained obvious. A further scan 3 months later showed new lesions in the cortex (D, axial T₂-weighted image, arrow).