Parental Mosaicism in “De Novo” Epileptic Encephalopathies

TO THE EDITOR

De novo disease-causing variants have been increasingly recognized in apparently sporadic, severe neurologic disorders in children, including developmental and epileptic encephalopathies¹ and autism.² Geneticists indicate that the risk of recurrence of these disorders in families with one affected child is approximately 1%; this accounts for the fact that one parent may have gonadal mosaicism.² In families with an affected child, the actual risk of recurrence may be as high as 50%.

Using single-molecule molecular inversion probes,³ we investigated the frequency of low-level parental mosaicism in somatic tissue obtained from parents and their affected children with an apparently de novo pathogenic variant (according to American College of Medical Genetics and Genomics criteria⁴) in 1 of 33 genes known to cause developmental and epileptic encephalopathies. Of 154 consecutively ascertained family trios (consisting of a child and his or her biologic parents), 123 (79.9%) yielded a minimum of 200 discrete captures (i.e., molecules) (see the Supplementary Methods section in the Supplementary Appendix, available with the full text of this letter at NEJM.org) in each parent; this coverage threshold provided 86.6% power to detect a 1% minor allele frequency (as calculated with the use of the binomial test). The variant was validated in each proband; paternity and maternity were genetically confirmed. Although ascertainment bias is possible, particularly in families with two affected children, genetic testing was commenced before the second affected child was born or before the child became clinically affected by the disorder. Three probands showed somatic mosaicism and so were excluded from the analysis.

We tested somatic tissue (blood or saliva) obtained from the parents in the remaining 120 families to infer gonadal mosaicism; of these, 10 parents (8.3%; 95% confidence interval, 3.4 to 13.3) had mosaicism for their child’s pathogenic variant (6 fathers and 4 mothers; minor allele frequency, 1.4 to 30.6%; mean, 12.9%; median, 9.4%) (Table 1). The minor allele frequency was well below that traditionally detected by means of Sanger sequencing in 8 of these 10 parents. In the saliva and blood samples obtained from 8 of the 10 parents with mosaicism (Table 1), the mutant allele had a similar frequency. Pathogenic variants occurred in eight genes. These genes included SCN1A in 3 of 40 families with apparently de novo SCN1A mutations; these findings showing that approximately 10% of children with an apparently de novo SCN1A variant had a parent with mosaicism replicated those of another study.⁵ In addition, one variant occurred in each of the following genes: SCN8A, GNB1, SLC6A1, DNM1, KCNT1, CACNA1A, and KCNQ2. Owing to the small sample size, we were unable to determine whether certain genes, such as those encoding ion channels, were more prone to mosaicism.

Table 1.

Parental Mosaicism in 8.3% of Parents Who Had a Child with a Diagnosis of an Apparently De Novo Monogenic Developmental and Epileptic Encephalopathy.^☆

Family No.	Proband Phenotype	Mutation	Parent with Mosaicism	Sample Type, % Mosaicism	Phenotype of Parent with Mosaicism	No. of Affected Siblings	Phenotype of Affected Sibling, Mutation Status
1	Dravet syndrome	SCN1A p.R101W	Father	Blood, 29.6; saliva, 16.7	Unaffected	0	NA
2	Dravet syndrome	SCN1A p.S1516☆	Mother	Blood, 17.6	Febrile seizure	0	NA
3	Dravet syndrome	SCN1A p.I1483Mfs☆18	Father	Blood, 30.6; saliva, 24.1	Febrile seizure	1	Dravet syndrome, SCN1A heterozygote
4	Developmental and epileptic encephalopathy	SCN8A p.L1331V	Father	Blood, 12.0; saliva, 4.7	Febrile seizures plus†	1	Febrile seizures plus,† focal seizures, learning difficulties, SCN8A heterozygote
5	Developmental and epileptic encephalopathy	KCNT1 p.R950Q	Father	Blood, 10.8; saliva, 14.0	Mild autosomal dominant nocturnal frontal lobe epilepsy	0	NA
6	Developmental and epileptic encephalopathy	KCNQ2 p.V567D	Mother	Blood, 1.4; saliva, 1.5	Unaffected	0	NA
7	Epilepsy with myoclonic–atonic seizures	SLC6A1 p.A334P	Mother	Blood, 8.0; saliva, 9.3	Unaffected	0	NA
8	Developmental and epileptic encephalopathy	GNB1 p.A326T	Father	Blood, 7.8	Unaffected	0	NA
9	Developmental and epileptic encephalopathy	CACNA1A p.A713T	Mother	Blood, 6.4; saliva, 8.5	Unaffected	1	Developmental and epileptic encephalopathy, CACNA1A heterozygote
10	Developmental and epileptic encephalopathy	DNM1 p.R237W	Father	Blood, 4.5; saliva, 4.1	Unaffected	1	Developmental and epileptic encephalopathy, DNM1 heterozygote
11	Developmental and epileptic encephalopathy	SYNGAP1 p.L150Vfs☆6	Inferred	ND	Unaffected	1	Developmental and epileptic encephalopathy, SYNGAP1 heterozygote
12	Epilepsy with myoclonic– atonic seizures	KIAA2022 p.R322☆	Inferred	ND	Unaffected	1	Epilepsy with myoclonic or atonic seizures, KIAA2022 heterozygote

^☆NA denotes not applicable, and ND not detected.

^†Febrile seizures plus are febrile seizures that occur after the age when these seizures usually occur (3 months to 6 years) or when there are concurrent afebrile tonic–clonic seizures.

In 13 of 120 families, a second child had seizures or a neurodevelopmental abnormality. In 5 of these 13 families, the affected sibling had a phenotype concordant with that of the proband and shared the proband’s mutation. However, parental mosaicism was detected in only 3 of these 5 families (these 3 families were captured in the 10 in which we observed parental mosaicism). Mosaicism in a parent of the other 2 families may have been below the level of detection by means of single-molecule molecular inversion probes or confined to gonadal tissue (which we did not test). If so, we have underestimated the true frequency of mosaicism in the parents. Conversely, only 1 of 8 siblings with a milder (discordant) phenotype carried their sibling’s mutation; mosaicism was detected in their father. Targeted high-coverage testing of parents who have a child with a developmental and epileptic encephalopathy due to an apparently de novo mutation may be helpful in counseling parents regarding the risk of recurrence.

A parental history of seizures was associated with an increased likelihood of parental mosaicism (P = 0.03 by Fisher’s exact test). Of the 16 parents who had a history of seizures, 4 had mosaicism and 12 did not; however, only 6 of 104 families with unaffected parents carried a variant that was also present, in a mosaic pattern, in either the mother or father (Table 1, and Fig. S1 in the Supplementary Appendix). The level of mosaicism correlated broadly with the severity of disease in the 4 affected parents who were found to have mosaicism.