Abstract
Elevated plasma levels of lipoprotein(a) [Lp(a)] are a significant genetic risk factor for the development of cardiovascular disease. Plasma Lp(a) levels are primarily modulated via a unique size polymorphism within the LPA gene, which results from copy number variation of the kringle IV2 (KIV2) domain. Despite a strong, inverse relationship between KIV2 size and plasma Lp(a), similarly sized alleles still associate with a wide range of Lp(a) levels. A number of single nucleotide polymorphisms (SNPs) within the LPA gene have been shown to further alter plasma Lp(a). The highly repeated nature of the KIV2 domain has posed significant challenges in sequencing and variant calling within the domain, and as such this region, unlike the rest of the gene, has remained largely uncharacterised for SNPs.
Next generation sequencing was combined with a tailored bioinformatics approach designed to characterise variation in the KIV2 domain of the LPA gene. This pipeline was used to screen for variation in 48 New Zealand European samples. Sequencing revealed 19 coding SNPs; the highest level of reported variation in the region to date. This number included 10 novel SNPs. Interestingly, many of the variants appeared to be present in single KIV2 copies. A number of the identified SNPs were predicted to be damaging to the kringle structure and were identified in samples exhibiting null Lp(a) phenotypes, making them good candidates for modulators of Lp(a).
The non-repeated regions of LPA were also screened for variation, returning three novel and 19 previously reported variants. Of particular interest was rs41259144, a non-synonymous SNP within KIV4. The variant was identified in four samples exhibiting a null Lp(a) phenotype and was predicted via protein modeling to be damaging to the kringle structure. Genotyping of the SNP within two New Zealand and two international cohorts revealed a significant association of the allele with lowered Lp(a) levels in European populations.
The high sensitivity of the method developed in this study enabled an in-depth assessment of KIV2 subtypes. According to currently published literature the KIV2 repeats can be classified into three subtypes based on the combinations of three synonymous SNPs within exon one. We identified five novel subtypes, characterised by unique combinations of these same SNPs. This finding has implications pertaining to the evolution of the domain. We hypothesize that the KIV2 domain is evolving as a functional unit as opposed to independent repeats and propose a mechanism whereby variation is homogenized across the domain via concerted evolution.
Next generation sequencing, when paired with the tailored bioinformatic approach developed in this thesis, represents a powerful method to detect variation in highly repeated regions of DNA and could be applied to other regions of the genome containing repetitive sequences. Application of this pipeline to LPA revealed a plethora of novel variation within KIV2 indicating that there is a larger reservoir of variation present in this region than currently reported. These results have implications in elucidating the contribution of variation within LPA to altered Lp(a) levels and cardiovascular disease risk. A complete understanding of the genetic basis of risk could enable early detection of high-risk individuals, allowing for adequate monitoring and treatment practices.