Novel NGS-Based Estimation of Kringle-IV-2 Repeats of LPA Gene Improves Genetic Identification of Elevated Lp(a) Levels

Abstract

<h3>Lead Author's Financial Disclosures</h3> M.R. is an employee of GB Healthwatch. <h3>Study Funding</h3> None. <h3>Background/Synopsis</h3> Lp(a) levels are genetically determined. Short isoforms within the variable Kringle-IV, type 2 (KIV-2) region of the LPA gene (that encodes for the apolipoprotein(a), apo(a) protein) is the single greatest genetic determinant of high Lp(a). Methods to directly assess this polymorphism may have clinical utility but is expensive and difficult to scale. Previous reporting indicates that 20-50% of people with short apo(a) isoforms do not have a corresponding tagging single nucleotide polymorphism (SNP). As next-generation sequencing (NGS) has become the standard platform for analyzing pathogenic variants causal of monogenic diseases and polygenic risk scores in cardiometabolic disease risk analysis, it would be advantageous to assess genetic risk for elevated Lp(a) within the same NGS-based test. We developed an NGS method to quantify KIV-2 repeats. <h3>Objective/Purpose</h3> To validate this metric for estimating KIV-2 repeat length (referred to as GB-KIV-2) directly from sequencing reads. <h3>Methods</h3> We performed NGS analysis on 97 patients with known Lp(a) levels from 2 academic clinics based in the United States. <h3>Results</h3> Random forest regression analysis of known SNPs that influence Lp(a) along with the GB-KIV-2 repeat metric indicate that the latter is the most important predictor of Lp(a) levels in this cohort (figure 1). We show a stepwise, inverse, linear relationship between the GB-KIV-2 metric and Lp(a) levels, Pearson correlation coefficient, r=-0.59, p-value: 2.0e-10. We compared the GB-KIV-2 with the gold standard immunoblotting method for isoform size and report a correlation coefficient, r=0.63, p-value =3.3e-4 The validation results show that GBinsight's NGS-based prediction for elevated Lp(a) has a sensitivity=82%, specificity=87% and accuracy=84%. <h3>Conclusions</h3> We conclude that inclusion of the GB-KIV-2 metric increases detection of the genetic causes of high Lp(a) while not increasing false positive rates. High Lp(a) may phenocopy familial hypercholesterolemia (FH). Genetic sequencing for FH confirmation and cascade screening is increasingly performed clinically. Including Lp(a) genetic screening may support differential diagnoses. FH). Genetic sequencing for FH confirmation and cascade screening is increasingly performed clinically. Including Lp(a) genetic screening may support differential diagnoses.