Migraine-Associated Common Genetic Variants Confer Greater Risk of Posterior vs. Anterior Circulation Ischemic Stroke☆

Petrea Frid, Huichun Xu, Braxton D. Mitchell, Mattias Drake, Johan Wassélius, Brady Gaynor, Kathleen A. Ryan, Anne‐Katrin Giese, Markus D. Schirmer, Kathleen Donahue, Robert E. Irie, Mark J.r.j. Bouts, Elissa C. McIntosh, Steven J. T. Mocking, Adrian V. Dalca, Eva Giralt‐Steinhauer, Lukas Holmegaard, Katarina Jood, Jaume Roquer, John W. Cole, Patrick F. McArdle, Joseph P. Broderick, Jordi Jiménez-Conde, Christina Jern, Brett Kissela, Dawn Kleindorfer, Robin Lemmens, James F. Meschia, Jonathan Rosand, Tatjana Rundek, Ralph L. Sacco, Reinhold Schmidt, Pankaj Sharma, Agnieszka Słowik, Vincent Thijs, Daniel Woo, Bradford B. Worrall, Steven J. Kittner, Jesper Petersson, Polina Golland, Ona Wu, Natalia S. Rost, Arne LindgrenJournal of Stroke and Cerebrovascular Diseases 2022 · DOI

TL;DR

This study investigates the genetic links between migraine and different types of ischemic stroke. Researchers found that genetic risk factors for migraine are more strongly associated with posterior circulation ischemic stroke (PCiS) than anterior circulation ischemic stroke (ACiS), suggesting a shared genetic predisposition. This finding could lead to a better understanding of stroke mechanisms and potentially new preventative strategies.

What Problem Does This Paper Address?

Ischemic stroke, a major cause of disability and death, can be broadly categorized into anterior circulation ischemic stroke (ACiS) and posterior circulation ischemic stroke (PCiS) based on the affected brain region. While migraine, particularly migraine with aura, is known to increase the risk of cardiovascular events and stroke, the specific relationship between migraine and these distinct stroke subtypes has not been thoroughly explored from a genetic perspective. Previous studies have hinted at a higher prevalence of infarct-like lesions in the posterior circulation territory in migraineurs, but the underlying genetic mechanisms remain unclear. This paper aims to bridge this gap by examining whether genetic variants associated with migraine also confer a differential risk for PCiS versus ACiS.

Methods

The researchers employed a polygenic risk score (PRS) approach. They utilized publicly available genome-wide association study (GWAS) data for three migraine phenotypes: any migraine, migraine without aura, and migraine with aura. These PRS were then calculated for individuals within a large ischemic stroke cohort (MRI-GENIE) that had been phenotyped for PCiS or ACiS based on MRI-DWI lesion location.

The core methodology involved:

  • Phenotyping: Ischemic stroke cases were meticulously classified as PCiS or ACiS using MRI-DWI images, with phenotyping performed by blinded researchers. Exclusion criteria were applied to ensure the purity of the stroke subtypes.
  • Polygenic Risk Score (PRS) Calculation: For each migraine phenotype, PRS were generated using summary statistics from a large migraine GWAS. This involved selecting single nucleotide polymorphisms (SNPs) associated with migraine at various p-value thresholds (e.g., 10⁻⁵ to 10⁻⁸) and weighting them by their effect sizes. The PRS for an individual is the sum of these weighted SNP effects.
  • Statistical Analysis:
    • Primary Analysis: Mean PRSs were compared between PCiS and ACiS patients to assess if migraine genetic risk was differentially associated with these stroke subtypes.
  • Secondary Analysis: PRSs were compared between PCiS patients and non-stroke control subjects, and separately between ACiS patients and non-stroke control subjects.
  • Transethnic Analysis: The study also included individuals of African ancestry to assess the generalizability of findings.
  • Statistical Controls: Analyses controlled for age, sex, and principal components of genetic ancestry. Bonferroni correction was applied for multiple testing in the primary analyses.

Datasets

  • Migraine GWAS Data: Publicly available summary statistics from a large meta-analysis of 22 migraine GWAS studies, involving approximately 375,000 individuals (59,674 cases and 316,078 controls) of European genetic ancestry.
  • Ischemic Stroke Cohort (MRI-GENIE): This cohort comprised patients with acute ischemic stroke lesions confirmed by MRI-DWI. After exclusions and quality control, the study included 505 PCiS patients and 1182 ACiS patients of European ancestry for the primary analyses. A smaller subset of 144 patients of African ancestry was included in transethnic analyses.
  • Non-Stroke Control Subjects: Genotype data from 15,396 non-stroke control subjects of European ancestry were used for comparison.

Key Insights and Ideas

  • Migraine's Genetic Footprint on Stroke Subtypes: The most striking insight is that genetic predisposition to migraine is not uniformly associated with all types of ischemic stroke. Instead, it appears to have a more pronounced genetic link with PCiS compared to ACiS. This suggests that distinct biological pathways might be involved in the pathogenesis of these stroke subtypes, and that migraine shares some of these pathways, particularly with PCiS.
  • Migraine Without Aura as a Key Driver: The association between migraine genetic risk and PCiS was particularly strong for the "migraine without aura" phenotype. This specificity is important, as it points towards particular aspects of migraine pathophysiology that might be relevant to stroke risk.
  • Potential for Shared Mechanisms: The findings strongly suggest shared genetic underpinnings, possibly related to dysregulation of cerebral vessel endothelial function or vascular smooth muscle function, which could contribute to both migraine and PCiS.
  • The Value of Phenotypic Specificity: This study highlights the importance of precisely phenotyping stroke subtypes (PCiS vs. ACiS) when investigating genetic associations. Treating all ischemic strokes as a single entity might obscure important subtype-specific genetic relationships.

Key Results

  • Primary Analyses (PCiS vs. ACiS):
    • Polygenic risk scores (PRSs) for "any migraine" were significantly higher in PCiS patients compared to ACiS patients across multiple p-value thresholds (p-range 0.001–0.010).
  • PRSs for "migraine without aura" were also significantly higher in PCiS compared to ACiS (nominal p-values 0.032–0.048), though these did not meet the strict Bonferroni corrected cutoff.
  • PRSs for "migraine with aura" did not show a significant differential association between PCiS and ACiS.
  • Secondary Analyses (Stroke vs. Controls):
    • "Any migraine" PRSs were associated with an increased risk of PCiS (p=0.011–0.03) and a decreased risk of ACiS (p=0.010–0.039) when compared to non-stroke controls.
  • "Migraine without aura" PRSs were significantly associated with an increased risk of PCiS (p=0.008–0.028) but showed no significant association with ACiS risk.
  • "Migraine with aura" PRSs showed no significant association with either PCiS or ACiS risk when compared to controls.
  • Transethnic Analyses: The findings for "any migraine" and "migraine without aura" remained largely consistent when including individuals of African ancestry, though the association for "migraine without aura" reached only borderline significance.

Limitations and Future Work

  • Selection Bias: The study relied on patients who underwent MRI and had visible DWI lesions, which might exclude individuals with less severe strokes or those not undergoing MRI, potentially introducing selection bias.
  • Migraine Comorbidity Data: Migraine comorbidity data within the ischemic stroke cohort were unavailable, preventing direct assessment of phenotypic associations.
  • Sample Size: While substantial, the sample size for specific stroke subtypes might have limited the power to detect subtle associations or perform stratified analyses for rarer stroke subtypes.
  • Correction for Multiple Testing: The authors acknowledge that the sample size precluded stringent correction for multiple testing in all analyses, particularly for the PRS method.
  • Future Work: The authors suggest that replication in larger studies with well-defined neuroimaging subtypes of ischemic stroke is warranted. Further research could explore the specific biological mechanisms, such as endothelial dysfunction, that link migraine and PCiS. Investigating the role of rare genetic variants and exploring diverse ancestries in both migraine and stroke cohorts are also important future directions.

Summary updated

Abstract

Our results suggest a stronger genetic overlap between unspecified migraine and migraine without aura with PCiS compared to ACiS. Possible shared mechanisms include dysregulation of cerebral vessel endothelial function.
BibTeX
@article{frid2022migraine,
  author = {Petrea Frid and Huichun Xu and Braxton D. Mitchell and Mattias Drake and Johan Wassélius and Brady Gaynor and Kathleen A. Ryan and Anne‐Katrin Giese and Markus D. Schirmer and Kathleen Donahue and Robert E. Irie and Mark J.R.J. Bouts and Elissa C. McIntosh and Steven J. T. Mocking and Adrian V. Dalca and Eva Giralt‐Steinhauer and Lukas Holmegaard and Katarina Jood and Jaume Roquer and John W. Cole and Patrick F. McArdle and Joseph P. Broderick and Jordi Jiménez-Conde and Christina Jern and Brett Kissela and Dawn Kleindorfer and Robin Lemmens and James F. Meschia and Jonathan Rosand and Tatjana Rundek and Ralph L. Sacco and Reinhold Schmidt and Pankaj Sharma and Agnieszka Słowik and Vincent Thijs and Daniel Woo and Bradford B. Worrall and Steven J. Kittner and Jesper Petersson and Polina Golland and Ona Wu and Natalia S. Rost and Arne Lindgren},
  title = {{Migraine-Associated Common Genetic Variants Confer Greater Risk of Posterior vs. Anterior Circulation Ischemic Stroke☆}},
  journal = {Journal of Stroke and Cerebrovascular Diseases},
  year = {2022},
  doi = {10.1016/j.jstrokecerebrovasdis.2022.106546},
}
Figure