Genetics of nonlesional focal epilepsy in adults and surgical implications

Nonlesional focal epilepsies (nlFE) are characterized by focal seizures, focal interictal epileptic discharges (IEDs), and the absence of epileptogenic lesions on magnetic resonance imaging (MRI). Sometimes nlFE are also described as nonacquired focal epilepsies (NAFE) as opposed to acquired, structural epilepsies, e.g., after cerebral ischemia, hemorrhage, or trauma. Nonlesional focal epilepsies account for 20–40% of all epilepsies [22] and they encompass a wide range of epilepsy syndromes, ranging from self-limited epilepsies in neonates and infants associated with distinct epilepsy genes (e.g., self-limited neonatal epilepsy, SeLNE), self-limited focal epilepsies with presumed complex inheritance in older children (e.g., self-limited epilepsy with centrotemporal spikes, SeLECTS or formerly Rolandic epilepsy), to defined genetic syndromes that begin at a variable age (e.g., epilepsy with auditory features, EAF); see Fig. 1. Yet, the majority of nlFE are isolated and account for a large share of patients seen in daily practice. More than 50% of focal epilepsies do not show structural abnormalities on routine magnetic resonance imaging (MRI; [8]). They can be at best classified by their seizure onset zone while their etiology often remains opaque. In this review, we focus on nlFE that occur or persist in adulthood. For self-limited focal epilepsies of infancy and childhood, overview articles can be found elsewhere [36, 49].

With the advent of next-generation sequencing and advances in bioinformatics and statistical genetics, the genetic basis of nlFE was progressively unearthed. Monogenic forms that can be explained by a single, causative gene and feature classic Mendelian pedigrees are the exception. Brain somatic gene variants, i.e., variants present only in local brain tissue, have been found in MRI-negative cases in resective specimens from epilepsy surgery and are associated with microscopic changes of cortical development. The larger share of nlFE, however, seems to rely on the interplay of multiple genetic variants in analogy to idiopathic generalized epilepsies (IGE) and other complex genetic diseases (e.g., schizophrenia or type 2 diabetes; [7, 10]). Yet unlike IGE, which display a more homogeneous phenotype and are firmly established as a polygenic disorder, etiologies in nlFE are presumably more heterogeneous and stretch beyond genetics. Unrecognized autoimmune inflammation, microscopic structural anomalies, and neurodegenerative changes in older individuals offer potential explanations. A positive family history or the presence of certain comorbidities should prompt genetic testing. The indication and consequences of genetic testing in the setting of presurgical evaluation are currently debated but viewed overall favorably.

Interestingly, while the advancements in DNA sequencing technologies heralded an era of gene discoveries for developmental and epileptic encephalopathies (DEEs), the first epilepsy gene to be identified was CHRNA4. It was detected in a large family with sleep-related frontal lobe seizures [37]. Nonetheless, familial focal epilepsy syndromes are rare. They usually show autosomal-dominant inheritance. But family history may be sparse or not available, and reduced penetrance can further complicate the recognition of a positive family history. Furthermore, seizures in other family members might have been overlooked (e.g., in the case of exclusively nocturnal seizures) or misinterpreted (e.g., in the case of focal aware seizures such as déjà vu auras). Therefore, seizure semiologies associated with familial epilepsies, such as nocturnal hypermotor seizures or focal aware seizures with auditory features, should give rise to a thorough review of the family history [30]. In most cases, affected individuals have normal intellect, normal neurological examination results, and seizures can usually be controlled well [23].

The most current classification differentiates four genetic focal epilepsy syndromes; each can be related to genetic variation in several genes ([30]; see also Table 1).

Malformations of cortical development (MCD) that range from regional disturbances of cortical architecture such as FCD to complex, pan-cerebral lesions such as hemimegalencephaly have been shown to be founded on genetic factors [12]. For FCD, however, germline variants appear to play a minor role [2, 34]. Somatic variants in various genes of the mTOR pathway (DEPDC5, MTOR, NPRL2/3) were identified in postsurgical resection tissue in structural lesions [2, 19, 44]. Also, more extended lesions harbor somatic variants in mTOR pathway genes, such as MTOR itself, AKT, and PIK3CA [13]. Of late, with SLC35A2, a “non-MTOR” gene was shown to be related to malformations of cortical development with oligodendroglial hyperplasia in epilepsy (MOGHE), a milder form of MCD [4]. Moreover, for various mTOR pathway genes, such as TSC1/TSC2 (causing tuberous sclerosis) and DEPDC5, a double-hit hypothesis has been postulated. Here, the combined effect of a germline variant in conjunction with brain somatic variants is thought to give rise to circumscribed MCD [28, 29]. Unlike germline variants, brain somatic variants arise during cortical development resulting in somatic mosaicism. Depending on the time of occurrence, variants can be limited to small fractions of brain cells. By definition, somatic variants are not detectable in leukocyte-derived DNA samples.

The observation that often FCD are not recognizable in presurgical MRI [27] gives rise to the question of whether a share of nlFE can also be explained by somatic variants in “hidden” FCD. In a series of nlFE patients who underwent epilepsy surgery, somatic variants in SLC35A2 were found in 17% of cases [45]. Interestingly, in two of the three reported patients, the pathologic work-up revealed FCD1a. Ongoing research will probably uncover further genes associated with brain somatic variants and nlFE [9]. Developments in structural MRI and postprocessing techniques will help identify more locally confined lesions in the future [41] and thereby render a share of today’s MRI-negative epilepsies “MRI-positive.”

In analogy to IGE, nlFE have been the focus of international, large-scale sequencing and genotyping consortia. Like IGE, nlFE have been shown to rely on complex genetic architecture, albeit to a lesser degree. The role of common genetic variants, i.e., variants that occur with a minor allele frequency of > 1% in the population, was established through genome-wide association studies (GWAS; [10]). These GWAS assess the association of single-nucleotide polymorphisms (SNPs) with complex traits. Since the effect size of each SNP by itself is very low, large cohorts are needed to capture significant associations. The latest and to date largest GWAS for epilepsies, including more than 29,000 individuals with epilepsy [11], showed an SNP-based heritability for nlFE of approximately 16%. In comparison to IGE, which features an SNP-based heritability for nlFE of approximately 40%, this effect appears to be rather moderate. Moreover, while the study identified 19 genome-wide significant loci for IGE, none was identified for nlFE. These data suggest a minor role of common variants for nlFE but could also mirror the higher heterogeneity of the nlFE cohort, potentially including mislabeled cases of acquired focal epilepsy, cases with brain somatic mutations, or unrecognized cases of autoimmune epilepsy that could have diluted the effect. Polygenic risk score (PRS) analyses, which estimate the aggregated effect of all SNPs weighted by their effect size on an individual level, underlined that common variants convey an increased risk for nlFE that is, however, smaller than in IGE [17, 21]. Large-scale analyses of exome sequencing data demonstrated that also ultra-rare missense variants (URVs) and truncating variants were enriched in nlFE [7], albeit less than in IGE. Truncating variants in genes, associated with familial nlFE, such as DEPDC5, appear to be enriched in nonfamilial nlFE [46]. Interestingly, enrichment patterns of URVs seem to differ between nlFE and IGE: While IGE exhibited a higher burden of URVs in gene sets derived from inhibitory neurons, nlFE carried a higher burden of URVs derived from excitatory neurons [14].

Genetic testing has become an established component of the diagnostic work-up of individuals with epilepsy. Decreasing costs for genetic testing and broader financial coverage by insurance companies made testing more widely available. The ascertainment of a genetic diagnosis enables genetic counseling, may give diagnostic and therapeutic guidance, and can help avoid further, potentially detrimental diagnostic tests [43]. Genetic testing should be considered for individuals with a high pretest probability. In the case of nlFE, testing should be performed of individuals with a positive family history of epilepsy, especially if the symptoms point to a specific, familial nlFE syndrome, such as SHE or EAF. Since families are often small, family members may not be contactable, and penetrance is usually incomplete, specific attention should be directed to seizure semiologies that are suggestive of familial nlFE syndromes, such as nocturnal hypermotor seizures or auditory, focal aware seizures. In nonfamilial nlFE, genetic testing should be considered in individuals who present with additional symptoms such as intellectual impairment, autism spectrum disorders, or dysmorphic features [15]. In patients with epilepsy and intellectual impairment (IQ < 70), the diagnostic yield of genetic testing can be up to 50% [47]. In this cohort, besides missense variants, also copy number variants, i.e., deletions or duplications of large DNA segments, account for about one third of positive cases.

If genetic testing is performed, either whole-exome sequencing (WES) or whole-genome sequencing (WGS) should be favored. Epilepsy gene panels that contain a curated list of known epilepsy-associated genes are obsolete since WES and WGS deliver a higher diagnostic yield (up to 45/48% for WES/WGS vs. 25% for panel diagnostics; [31, 35]). If available, WES/WGS should be performed as a trio analysis, i.e., including both parents for the sake of interpretation of unknown variants WES is today the standard in most diagnostic laboratories, and analysis pipelines usually include CNV analyses, which render chromosomal microarrays redundant. See Table 2 for an overview of indications for genetic testing in nlFE. General recommendations for genetic testing in individuals with epilepsy in Germany are published and regularly reviewed by the Epilepsy and Genetics Commission (Kommission Epilepsie und Genetik) of the German ILAE branch (DGfE; Dt. Ges. für Epileptologie; http://​www.​dgfe.​org/​home/​index,id,528.​html) and have recently been published by the ILAE genetics commission [15].

Resective epilepsy surgery for individuals with genetic epilepsies may seem counterintuitive at first glance. Yet, as in tuberous sclerosis, resective surgery has been established for many years [48]. Increasing knowledge about the associations of mTOR-pathway genes beyond TSC1 and TSC2 and their association with focal lesional as well as nonlesional epilepsies has kindled the debate about whether patients with nlFE should systematically undergo genetic testing during presurgical work-up [20, 38]. A better understanding of the relation between genetic diagnosis and postoperative outcome could empower clinicians to make better predictions about the odds of successful epilepsy surgery. Systematic reviews and large case collections testify to the low chances of effective seizure control in patients carrying variants in genes related to channel function or synaptic transmission [5, 38]. The same studies observe far more promising results for variants in mTOR-pathway genes. This trend is corroborated by findings from a Dutch epilepsy center that also highlights the increasing use of genetic testing in nlFE [32]. Although clinicians should be wary of performing surgery in patients with channelopathies or synaptopathies, these individuals should not be categorically excluded from presurgical evaluation [42]. A recent survey by the DGfE among German epilepsy centers showed that genetic testing is viewed favorably in many case constellations and, in the case of nlFE, recommended by more than 90% of the survey respondents [5]. Founded on these results, the Epilepsy and Genetics Commission (DGfE) recommends genetic testing as part of the presurgical work-up.