IGF Gene-network Identifies Specific Interactions Underlying Extreme SBP in Developing Children
Parmar, P; Pennell, CE; Palmer, LJ; Briollais, L
MetadataShow full metadata
Genes of the Insulin-like Growth Factor (IGF) network have been shown to be associated with blood pressure and cardiovascular health. Here we aim to further current knowledge relating to the molecular mechanisms underlying the developmental origins of hypertension by accounting for some genetic complexity of Systolic Blood Pressure (SBP) by modelling the nine IGF-Axis genes (ligands; IGF1 and IGF2, receptors; IGF1R and IGF2R, and binding proteins; IGFBP1-IGFBP5) as a gene-network and provide insight for the early onset of hypertension in growing children. We applied the two-stage gene-network methodology from Tsonaka et al using longitudinal data on growth and SBP across childhood and adolescence (5-18 years) from the Western Australian Pregnancy and Birth Cohort (Raine) Study. For the first-stage analysis, a mixed-effects logistic model is used to specify the probability to carry at least one rare allele (dominant model) or two rare alleles (recessive model) as a Bernouilli trial and returns the empirical-Bays estimates of the random effects. The second stage analysis tests the gene-specific estimates using the empirical-Bayes estimates from the first stage as covariates in the mixed-effects model for longitudinal SBP. A linear mixed-effects model with random intercepts and random slopes, adjusting for age group (≤ 15 years and > 15 years), sex and BMI was used to create the longitudinal gene-network model. At a Bonferroni-corrected significance level of p<0.05, we identified 16 gene-gene interactions influencing the bottom 5% of recorded SBP measures and 15 gene-gene interactions for the extreme top 5% of SBP measures. Only four of these interactions were present in children aged ≤15 years, one in females (for increasing SBP in the bottom 5%) and three in males for increasing SBP in the top 5%. The remainder of interactions were present in children aged 15 years and above. The main interactions for females was due to the interaction between IGF1R and IGFBP4 (reducing SBP) and IGFBP2-IGFBP5 (increasing SBP), and for males was between IGF2R and IGFBP1 (reducing SBP) and IGF2-IGFBP2 (increasing SBP) across the top 5% of SBP measures. The results attained are reasonable and align logically with current literature, particularly as IGFBP1 has been linked with both positive and negative associations to SBP and to well-known risk factors of cardiovascular diseases. Polymorphisms within IGF2 have already been shown to influence regulation of blood pressure in obese children. We found that this gene-network is modified with age; this in itself may be due to a number of reasons including diet, hormones and developmental growth over time, particularly post-puberty. Further investigation using this method and other modern high dimensional statistical methods to analyse gene-networks in cohort and large consortium data with external biological knowledge (such as exonic, regulatory gene information and the influence of hormones) would be ideal to validate our findings and improve accuracy surrounding the estimates produced from these models and further improve the power to detect complex interactions. Through characterizing the association between multiple genes and disease outcomes we will offer new insight into disease aetiology whilst providing tools for making individualized treatment decisions.