Distinct gene-set burden patterns underlie common generalized and focal epilepsies

Abstract

Background: Analyses of few gene-sets in epilepsy showed a potential to unravel key disease associations. We set out to investigate the burden of ultra-rare variants (URVs) in a comprehensive range of biologically informed gene-sets presumed to be implicated in epileptogenesis.

Methods: The burden of 12 URV types in 92 gene-sets was compared between cases and controls using whole exome sequencing data from individuals of European descent with developmental and epileptic encephalopathies (DEE, n = 1,003), genetic generalized epilepsy (GGE, n = 3,064), or non-acquired focal epilepsy (NAFE, n = 3,522), collected by the Epi25 Collaborative, compared to 3,962 ancestry-matched controls.

Findings: Missense URVs in highly constrained regions were enriched in neuron-specific and developmental genes, whereas genes not expressed in brain were not affected. GGE featured a higher burden in gene-sets derived from inhibitory vs. excitatory neurons or associated receptors, whereas the opposite was found for NAFE, and DEE featured a burden in both. Top-ranked susceptibility genes from recent genome-wide association studies (GWAS) and gene-sets derived from generalized vs. focal epilepsies revealed specific enrichment patterns of URVs in GGE vs. NAFE.

Interpretation: Missense URVs affecting highly constrained sites differentially impact genes expressed in inhibitory vs. excitatory pathways in generalized vs. focal epilepsies. The excess of URVs in top-ranked GWAS risk-genes suggests a convergence of rare deleterious and common risk-variants in the pathogenesis of generalized and focal epilepsies.

Funding: DFG Research Unit FOR-2715 (Germany), FNR (Luxembourg), NHGRI (US), NHLBI (US), DAAD (Germany).

Keywords: Burden analysis; Epilepsy; Exome sequencing; Gene-sets; Ultra-rare variants.

Fig. 1

Exome-wide burden of ultra-rare variants in the epilepsies. The burden in developmental and epileptic encephalopathies (DEE), genetic generalized epilepsies (GGE) and non-acquired focal epilepsies (NAFE) in (A) 19,402 protein coding genes and (B) 1,743 genes with probability of loss-of-function intolerance (pLI) score > 0·995 is shown in multiple classes of variants (y-axis; see methods) as odds ratio (x-axis) from Likelihood Ratio Test (bars indicate 95% confidence intervals). False-Discovery-Rate-adjusted p values (synonymous variants analysis p values were not adjusted) are indicated with stars as follows: no star > 0·05, * < 0·05, ** < 0·005, *** < 0·0005, **** < 0·00005.

Fig. 2

Burden of ultra-rare missense variants in brain expressed and developmental genes. The burden of benign or damaging missense variants and missense variants in highly constrained sites in developmental and epileptic encephalopathies (DEE), genetic generalized epilepsies (GGE) and non-acquired focal epilepsies (NAFE) is shown in gene-sets based on levels of RNA/protein expression in the cortex and hippocampus (A) or enrichment in adult or developing brain (B). Gene-sets are shown on the y-axis (number of genes in parenthesis). Log odds ratio (Likelihood Ratio Test) are shown on the x-axis (error bars indicate 95% confidence intervals). The variant classes are shown in vertical panels. False-Discovery-Rate-adjusted p values are indicated with stars as follows: no star > 0·05, * < 0·05, ** < 0·005, *** < 0·0005, **** < 0·00005. High, medium and low expression categorization was based on expression levels in Gene Tissue Expression ​Project portal (GTEx). Brain-enriched (with more than four-fold expression in the brain compared to other tissues) and brain-enhanced genes (higher but less than four-fold expression) were obtained from the Human Protein Atlas.

Fig. 3

Burden in neuronal and glial cells, ion channels, receptors and related interactors. The burden in developmental and epileptic encephalopathies (DEE), genetic generalized epilepsies (GGE) and non-acquired focal epilepsies (NAFE) is shown on the x-axis (log-odds from Likelihood Ratio Test; error bars indicate 95% confidence intervals). Gene-sets are shown on the y-axis (number of genes in parenthesis). The variant classes are shown in vertical panels. False-Discovery-Rate-adjusted p values are indicated with stars as follows: no star > 0·05, * < 0·05, ** < 0·005, *** < 0·0005, **** < 0·00005. (A) Burden in genes enriched in specific brain cells including neuron- or glia-enriched genes and their subtypes. (B) Burden in key biologically informed neuronal gene-sets with known or suspected relation to epilepsy. NMDA: N-Methyl-Dextro-Aspartate. ARC: neuronal activity-regulated cytoskeleton-associated protein (interactors). PSD-95: Post-Synaptic-Density protein 95 (interactors).

Fig. 4

Enrichment in major neuronal synapses and pathways. Panels show comparison of enrichment patterns in developmental and epileptic encephalopathies (DEE), genetic generalized epilepsies (GGE) and non-acquired focal epilepsies (NAFE) in GABAergic and glutamatergic synapses and pathway genes based on (A) Gene-Ontology (GO) and (B) Kyoto Encyclopaedia for Genes and Genomes (KEGG). The burden is shown on the x-axis (log-odds from Likelihood Ratio Test; error bars indicate 95% confidence intervals). Gene-sets are shown on the y-axis (number of genes in parenthesis). The variant classes are shown in vertical panels. False-Discovery-Rate-adjusted p values are indicated with stars as follows: no star > 0·05, * < 0·05, ** < 0·005, *** < 0·0005, **** < 0·00005. Complete groups, genes specific to one of the two synapses/pathways as well as their intersection were tested.

Fig. 5

Burden of ultra-rare variants in groups of epilepsy-related known disease genes. The burden in five gene-sets (y-axis; number of genes in parenthesis) in developmental and epileptic encephalopathies (DEE), genetic generalized epilepsies (GGE) and non-acquired focal epilepsies (NAFE) (horizontal panel) in selected variant classes (vertical panels) is shown on the x-axis (log odd ratios from Likelihood Ratio Test; error bars indicate 95% confidence intervals). False-Discovery-Rate-adjusted p values are indicated with stars as follows: no star > 0·05, * < 0·05, ** < 0·005, *** < 0·0005, **** < 0·00005. NDD-Epilepsy: neurodevelopmental disorders with epilepsy. FMPR: Fragile-X Mental Retardation Protein. MGI: Mouse Genome Informatics database.

Fig. 6

Risk elements in GWAS top-ranked genes and co-expression modules. The burden of missense variants in highly constrained sites (log-odds on the x-axis; error bars indicate 95% confidence intervals) in developmental and epileptic encephalopathies (DEE), genetic generalized epilepsies (GGE) and non-acquired focal epilepsies (NAFE) is shown in gene-sets (y-axis; number of genes in parenthesis) representing (A) Generalized or Focal epilepsy (presumed monogenic) genes as well as top-ranked 100 genes from GWAS of generalized and focal epilepsies, and (B) co-expressed genes identified in post-mortem brain tissues of healthy individuals (module of 320 genes) or in brain tissues from TLE patients (network of 395 genes) as well as two sub-modules of this network (M1 and M2). False-Discovery-Rate-adjusted p values are indicated with stars as follows: no star > 0·05, * < 0·05, ** < 0 ·005, *** < 0·0005, **** < 0·00005.