Targeted sequencing of 351 candidate genes for epileptic encephalopathy in a large cohort of patients

Abstract

Background: Many genes are candidates for involvement in epileptic encephalopathy (EE) because one or a few possibly pathogenic variants have been found in patients, but insufficient genetic or functional evidence exists for a definite annotation.

Methods: To increase the number of validated EE genes, we sequenced 26 known and 351 candidate genes for EE in 360 patients. Variants in 25 genes known to be involved in EE or related phenotypes were followed up in 41 patients. We prioritized the candidate genes, and followed up 31 variants in this prioritized subset of candidate genes.

Results: Twenty-nine genotypes in known genes for EE (19) or related diseases (10), dominant as well as recessive or X-linked, were classified as likely pathogenic variants. Among those, likely pathogenic de novo variants were found in EE genes that act dominantly, including the recently identified genes EEF1A2, KCNB1 and the X-linked gene IQSEC2. A de novo frameshift variant in candidate gene HNRNPU was the only de novo variant found among the followed-up candidate genes, and the patient's phenotype was similar to a few recent publications.

Conclusion: Mutations in genes described in OMIM as, for example, intellectual disability gene can lead to phenotypes that get classified as EE in the clinic. We confirmed existing literature reports that de novo loss-of-function HNRNPUmutations lead to severe developmental delay and febrile seizures in the first year of life.

Keywords: De novo; HNRNPU; X‐linked; epileptic encephalopathy; loss‐of‐function; prioritization; recessive; targeted panel sequencing.

Figure 1

Raw CADD‐scores for novel unique variants in our patient cohort versus unique variants in the ExAC database in three gene categories. (In all figures: note that most indels get no CADD‐scores).

Figure 2

Raw CADD‐scores for novel variants in our patient cohort in known genes for EE or related phenotypes. After checking the variants in relatives, and considering the patient's phenotype, variants were classified into probably benign and probably pathogenic. Recessive‐ and dominant‐acting genes are shown separately.

Figure 3

Per gene: Excess frequency of novel variants in TEGA versus ExAC database (X‐axis) versus average CADD‐scores for novel variants in TEGA samples (Y‐axis). Genes are classified into known EE genes, genes for intellectual disability, or phenotypes that may present with seizures, and candidate genes.