Mendelian randomization analysis of blood uric acid and risk of preeclampsia: based on GWAS and eQTL data.

Abstract

Background: The causal association between blood uric acid and preeclampsia (Preeclampsia, PE) has not been conclusively established based on the literature reviewed to date. This bi-directional Mendelian randomization study aimed to investigate the bi-directional causal association between blood uric acid concentration and PE at different genetic levels.

Methods: Pooled data on preeclampsia (sample size = 82,085) and blood uric acid (sample size = 129,405) were conducted based on publicly available genome-wide association analysis (Genome-Wide Association Study, GWAS) on the East Asian populations regarding preeclampsia and blood uric acid, respectively. We assessed blood uric acid and PE associations using two-sample Mendelian randomization (TSMR) analyses based on GWAS pooled statistics using inverse variance weighted (Inverse variance weighted), MR-Egger, and Weighted median (Weighted median) to examine the association between blood uric acid and pre-eclampsia. Causal relationship between blood uric acid and pre-eclampsia.Cochran's Q statistic was used to quantify the heterogeneity of instrumental variables among other methods. Subsequently, we extracted the expression quantitative trait loci (eQCTL, Expression quantitative trait loci) data corresponding to each gene as the instrumental variables using the genes corresponding to the intersecting instrumental variables of the exposure and the outcome in the respective analyses of the forward and backward TSMR respectively, so as to analyze the genetic causality of the genes with the different forward and backward TSMR methods further. Inverse variance weighted (IVW) was used to analyze the genetic causality of genes with different positive and negative outcomes.

Results: Genetically determined blood uric acid level IVW method, ratio (OR) 1.30, 95% confidence interval (CI): [0.6, 2.83], p = 0.51 was not risk associated with PE. In addition according to the inverse MR analysis, we found an OR of 0.99, 95% CI [0.99, 1.0], p = 0.999) for PE on blood uric acid level IVW method and no significant heterogeneity in instrumental variables or level polytropy was found. (ii) Although GWAS data suggested no risk association between PE and uric acid, gene association analysis of eQTL data at blood uric acid levels with PE suggested a risk effect of the TP53INP1 gene for PE (IVW, OR = 11.476, 95% CI 2.511-52.452, p = 1.648 × 10^-3) and a protective effect of CTSZ (IVW, OR = 0.011, 95% CI 0.001-0.189, p = 1.804 × 10^-3), while a risk effect of ETV7 on hyperuricemia was suggested in a genetic association analysis of PE eQTL data with blood uric acid levels (OR = 1.018, 95% CI 1.007-1.029, p = 1.289 × 10^-3).

Conclusion: Our MR (Mendelian Randomization) study based on the GWAS database did not support a bidirectional causal effect between blood uric acid levels and PE, whereas MR based on quantitative trait loci suggested that TP53INP1, which affects uric acid levels, has a risk association for PE, whereas CTSZ is protective against preeclampsia. Among the genes affecting PE the ETV7 gene may play a positive role in elevating uric acid levels.