Abstract
Leu C, Avbersek A, Stevelink R, Custodio HM, Chen S, Speed D, Bennett CA, Jonsson L, Unnsteinsdóttir U, Jorgensen AL, Cavalleri GL, Delanty N, Craig JJ, Depondt C, Johnson MR, Koeleman BPC, Hassanin E, Omidvar ME, Krause R, Lerche H, Marson AG, O'Brien TJ, Sander JW, Sills GJ, Striano P, Zara F, Stefansson H, Stefansson K, May P, Neale BM, Lal D, Berkovic SF; Epi25 Collaborative; EpiPGX Consortium; Sisodiya SM. EBioMedicine. 2025; 115:105675. PMID: 40240269 Background: Epilepsy is one of the most common neurological disorders, affecting over 50 million people worldwide. One-third of people with epilepsy do not respond to currently available anti-seizure medications, constituting one of the most important problems in epilepsy. Little is known about the molecular pathology of drug resistance in epilepsy, in particular, possible underlying genetic factors are largely unknown. Methods: We performed a genome-wide association study (GWAS) in two epilepsy cohorts of European ancestry, comparing drug-resistant (N = 4208) to drug-responsive individuals (N = 2618), followed by meta-analyses across the studies. Next, we performed subanalyses split into two broad subtypes: acquired or non-acquired focal and genetic generalized epilepsy. Findings: Our drug-resistant versus drug-responsive epilepsy GWAS meta-analysis showed no significant loci when combining all epilepsy types. Sub-analyses on individuals with focal epilepsy (FE) identified a significant locus on chromosome 1q42.11-q42.12 (lead SNP: rs35915186, P = 1·51 × 10−8, OR[C] = 0·74). This locus was not associated with any epilepsy subtype in the latest epilepsy GWAS (lowest uncorrected P = 0·009 for FE vs healthy controls), and drug resistance in FE was not genetically correlated with susceptibility to FE itself. Seven genome-wide significant SNPs within this locus, encompassing the genes CNIH4, WDR26, and CNIH3, were identified to protect against drug-resistant FE. Further transcriptome-wide association studies (TWAS) imply significantly higher expression levels of CNIH3 and WDR26 in drug-resistant FE than in drug-responsive FE. CNIH3 is implicated in AMPA receptor assembly and function, while WDR26 haploinsufficiency is linked to intellectual disability and seizures. These findings suggest that CNIH3 and WDR26 may play a role in mediating drug response in focal epilepsy. Interpretation: We identified a contribution of common genetic variation to drug-resistant focal epilepsy. These findings provide insights into possible mechanisms underlying drug response variability in epilepsy, offering potential targets for personalised treatment approaches. Funding: This work is part of the European Union's Seventh Framework Programme (FP7/2007-2013) under grant agreement no. 279062 (EpiPGX) and the Centers for Common Disease Genomics (CCDG) program, funded by the National Human Genome Research Institute (NHGRI) and the National Heart, Lung, and Blood Institute (NHLBI).
Commentary
An enduring challenge in epilepsy treatment is the persistent problem of drug resistance. 1 Overcoming this issue has become, in many ways, a Sisyphean task. In Greek mythology, Sisyphus was condemned by Zeus to eternally push a massive boulder up a hill, only to watch it roll back down each time it neared the summit. The history of epilepsy drug treatments has followed a similarly futile pattern, with antiseizure medications (ASMs) acting as boulders that never quite crest the hill of consistent efficacy.
Despite decades of intensive research and the development of more than 25 ASMs, approximately one-third of people with epilepsy continue to experience uncontrolled seizures. 1 This condition, termed drug-resistant epilepsy (DRE), remains strikingly prevalent, and its incidence has remained unchanged for decades, even with the discovery of new drugs, underscoring the intractable nature of the problem. 2 The International League Against Epilepsy (ILAE) defines DRE as the “failure of adequate trials of two tolerated, appropriately chosen and used ASM schedules (whether as monotherapies or in combination) to achieve sustained seizure freedom.” 3
The biological underpinnings of DRE are complex and still poorly understood. Hypotheses have included drug targets, multidrug transporters, disease severity, epigenetic factors, and pathological brain networks. 4 However, none of these has yielded a unifying explanation, leaving a substantial proportion of people with epilepsy consigned to a lifetime of refractory seizures and medication side effects.
While the challenge of DRE remains unresolved, new research offers a promising clue: genetics. In a large collaborative genome-wide association study (GWAS), Leu et al 5 examined common single-nucleotide polymorphisms (SNPs) in 2 large datasets of epilepsy patients of Western or Central European ancestry. Participants were recruited from the EpiPGX Consortium, a multicenter pharmacogenetics initiative, and the Epi25 Collaborative, the largest sequencing study in epilepsy to date. The investigators compared individuals with DRE (n = 4208) to those who were drug-responsive (n = 2618), followed by meta-analyses. They also conducted subgroup analyses based on seizure onset type, focal or generalized epilepsy. Their work yielded 2 key insights: (1) drug resistance in focal epilepsy has a common genetic component and (2) this component localizes to a small region on chromosome 1 containing 3 candidate genes: CNIH4, WDR26, and CNIH3.
This discovery is significant. Previously, only 1 GWAS had attempted to identify genetic factors in DRE, but its small sample size (889 patients) lacked statistical power and failed to yield any genome-wide significant associations. 6 In contrast, Leu et al identified 7 SNPs in high linkage disequilibrium on chromosome 1 that were significantly associated with drug resistance in focal epilepsy. Linkage disequilibrium refers to genetic variants that are inherited together more often than expected by chance, a clue that the region may harbor variants with functional or disease-related significance. Fine mapping narrowed the signal to a 161 kb region encompassing the 3 candidate genes. Notably, transcriptome-wide association studies using brain tissue revealed increased expression of CNIH3 and WDR26 in drug-resistant focal epilepsy, along with 18 distinct splicing events mapping to these 3 candidate genes, strengthening the case for their potential functional relevance.
Each of the implicated genes offers a possible mechanistic link to drug resistance. CNIH3 is involved in α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptor assembly and function 7 ; WDR26 haploinsufficiency is associated with seizures and intellectual disability 8 ; and CNIH4, although expressed in the brain, is only distantly related to CNIH3 and lacks its AMPA receptor binding domain,9,10 making its role in epilepsy unclear. Importantly, the identified SNPs are not necessarily causal but instead point to a genomic region that likely harbors one or more causal variants. Identifying these variants and the pathways they disrupt remains a crucial next step to understanding the biology and pathophysiology of DRE.
To enhance the reliability of their findings, the authors applied a more stringent definition of DRE, classifying individuals as drug-resistant only if they had experienced 4 or more seizures per year despite appropriate treatment. This approach excluded patients with only 1–3 seizures annually, who, while considered drug-resistant under broader ILAE criteria, can introduce ambiguity into genetic studies due to the challenge of determining whether infrequent seizures reflect true treatment failure. Conversely, individuals were deemed drug-responsive if they had been seizure-free for at least 12 months. Although these stricter definitions reduced the sample size by up to 30%, they likely sharpened the contrast between drug-resistant and drug-responsive groups, improving the precision and reliability of downstream genetic analyses.
Still, several limitations should be acknowledged. First, the study population consisted exclusively of individuals of European ancestry, and it remains uncertain whether these findings generalize to more diverse populations. Second, only about 20% of participants (1257 individuals) had generalized epilepsy, limiting the power to detect associations in this subgroup. However, stratifying by epilepsy type proved valuable. While combining all epilepsy types yielded no significant associations, focusing on focal epilepsy uncovered a clear genetic signal. This suggests that the mechanisms underlying drug resistance differ between focal and generalized epilepsies, and that lumping them together can obscure meaningful subtype-specific associations.
Perhaps the most impactful contribution of this study is its proof of principle that common genetic variation contributes to drug resistance in focal epilepsy. The effect sizes observed are modest. For example, the lead SNP was associated with only a 26% reduction in the odds of developing drug resistance, likely too small to meaningfully guide individual treatment decisions. However, this discovery paves the way for developing a clinically useful polygenic risk score for DRE by aggregating the effects of multiple small-effect variants across additional independent risk loci to better predict drug response in people with epilepsy. Future studies of larger cohorts will also be needed to understand how these genetic markers interact with specific ASMs and comorbidities, which could further refine their clinical relevance.
In sum, this work represents a pivotal step forward. It provides foundational evidence that common genetic factors are involved in drug resistance in focal epilepsy and lays the groundwork for more personalized treatment strategies in the future. While not yet clinically actionable, the study marks an encouraging advance toward overcoming one of epilepsy's most persistent and frustrating challenges. As the Sisyphean task of beating DRE continues, this study offers renewed hope that the summit may finally be within reach and that the boulder might, at last, stay at the top.
