Genome-wide pleiotropy analysis of longitudinal blood pressure and harmonized cognitive performance measures

Abstract

Introduction: Identifying pleiotropy for blood pressure (BP) and cognitive performance measures may indicate mechanistic links between hypertension and Alzheimer's disease (AD).

Methods: We performed a pleiotropy genome-wide association study (GWAS) for paired measures of systolic, diastolic, pulse, and mean arterial pressure with memory, executive function, and language scores using 116,075 exam data from 25,726 participants in clinic-based and prospective cohorts. Significant findings were evaluated by Bayesian colocalization and differential gene expression in brain tissue from pathologically confirmed AD cases with and without clinical symptoms.

Results: Genome-wide significant pleiotropy for BP and cognitive performance with JPH2, GATA3, PAX2, LOC105371656, and SUFU in the total sample; RTN4, ULK2, SORBS2, and LOC100128993 in prospective cohorts; and ADAMTS3 and LINC02946 in clinic-based cohorts. Six pleiotropic loci influence cognition directly, and six genes were differentially expressed between pathologically confirmed AD cases with and without antemortem cognitive impairment.

Discussion: Our results provide insight into mechanisms underlying hypertension and AD.

Highlights: Genome-wide significant pleiotropy in blood pressure (BP) and cognitive performance measures were identified with 11 novel loci: JPH2, GATA3, PAX2, LOC105371656, SUFU in the total sample; RTN4, ULK2, SORBS2, LOC100128993 in prospective cohorts; and ADAMTS3, LINC02946 in clinic-based cohorts. SUFU, RTN4, SORBS2, ADAMTS3, and GATA3 affected cognition directly rather than through BP. ACTR1A, HIF1AN, ADAMTS3, RTN4, SORBS2, and SUFU at pleiotropic loci were differentially expressed among controls and pathologically confirmed AD cases with and without clinical symptoms.

Keywords: Alzheimer's disease; blood pressure; cognitive domain score; differential gene expression; genome‐wide association study; longitudinal study; pathway analysis; pleiotropy.

FIGURE 1

LocusZoom plots for top‐ranked pleiotropic loci emerging from the total sample. (A) rs6031436 and JPH2 variants showing pleiotropy between SBP and language. (B) rs6031436 and JPH2 variants showing pleiotropy between PP and language. (C) rs263419 and GATA3 variants showing pleiotropy between SBP and language. (D) rs7902306 and SUFU variants showing pleiotropy between DBP and language. A purple diamond indicates the top‐ranked SNP at each locus. SNPs are color‐coded according to their LD (r ²) with the top‐ranked SNP in the region. Horizontal dotted lines represent the GWS threshold of P = 5×10⁻⁸. Vertical blue lines indicate locations of the high recombination rate among SNPs at the chromosomal position. Approximate location, transcription direction, and coding portions (exons represented by vertical bars) of genes are shown below the diagram. cM, centimorgan; DBP, diastolic blood pressure; GWS, genome‐wide significant; LD, linkage disequilibrium; Mb, megabase; PP, pulse pressure; SBP, systolic blood pressure; SNP, single nucleotide polymorphism.

FIGURE 2

LocusZoom plots for top‐ranked pleiotropic loci emerging from the clinic‐based or prospective cohort samples. (A) rs157398 and ULK2 variants showing pleiotropy between DBP and executive function in the prospective cohorts. (B) rs10518102 and ADAMTS3 variants showing pleiotropy between PP and language in the clinic‐based cohorts. (C) rs6707036 and RTN4 variants showing pleiotropy between SBP and language in the prospective cohorts. (D) rs2168164 and SORBS2 variants showing pleiotropy between PP and memory in the prospective cohorts. A purple diamond indicates the top‐ranked SNP at each locus. SNPs are color‐coded according to their LD (r ²) with the top‐ranked SNP in the region. Horizontal dotted lines represent the GWS threshold of P = 5×10⁻⁸. Vertical blue lines indicate locations of the high recombination rate among SNPs at the chromosomal position. Approximate location, transcription direction, and coding portions (exons represented by vertical bars) of genes are shown below the diagram. cM, centimorgan; DBP, diastolic blood pressure; GWS, genome‐wide significant; Mb, megabase; PP, pulse pressure; SBP, systolic blood pressure; SNP, single nucleotide polymorphism.

FIGURE 3

Pleiotropic genes significantly differentially expressed in the DLPFC in pathologically diagnosed AD cases with and without clinical symptoms prior to death and cognitively normal controls. Genes shown are significantly differentially expressed in the DLPFC among pathologically confirmed symptomatic (SymAD) and asymptomatic (AsymAD) AD cases, and cognitively normal controls. Differential expression levels in the DLPFC were estimated and meta‐analyzed using the ROSMAP, BUADRC, and FHS datasets. (A) ACTR1A, (B) HIF1AN, (C) ADAMTS3, (D) RTN4, (E) SORBS2, and (F) SUFU. Units on the Y‐axis represent log‐transformed (log₂) expression of genes. *FDR < 0.05, **FDR < 0.01, ***FDR < 0.001. AD, Alzheimer's disease; BUADRC, Boston University Alzheimer's Disease Research Center; DLPFC, Dorsolateral Prefrontal Cortex; FDR, false discovery rate; FHS, Framingham Heart Study; ROSMAP, Religious Orders Study/Rush Memory and Aging Project.