A genome-wide approach to evaluate the genetic basis of lipoprotein(a)

Lipoprotein(a) [Lp(a)] is a lipoprotein consisting of a core LDL-like particle and the glycoprotein apolipoprotein(a) [apo(a)] attached. It is a quantitative genetic trait, mostly influenced by the size of the apo(a) isoforms, which are expressed by copy-number-variations in the LPA gene (Kringle-IV repeats). Family studies suggest that about 90% of variation in Lp(a) is heritable, with a major fraction attributable to either the Kringle-IV repeats or SNPs in the LPA gene cluster. So far, genome-wide association studies on Lp(a) values only revealed significantly associated SNPs within this gene cluster. However, these studies were limited by either small sample size, lack of replication samples or by a focus on certain subgroups. It has also been shown, that SNPs in this gene region correlate with the apo(a) isoforms and that association with Lp(a) might be partly due to this correlation. Since measurements of apo(a) isoforms are very elaborative and time-consuming, these studies were limited by small sample sizes, leading to inconsistent and non-generalizable results. We aim to investigate the genetic regulation of Lp(a) via a genome-wide approach in seven discovery (n~21500) and three replication studies (n~4400). In all samples, apo(a) isoform measurements are or will be available and will be adjusted for in the genome-wide analyses. This approach improves the power to detect SNPs in new gene regions by reducing the residual variance and by excluding the association signals that are only due to correlation with apo(a) isoforms. In a first preliminary meta-analysis of genome-wide data in three studies (KORA F3, KORA F4, FamHS, n~7900), especially one SNP in the LPA gene region was shown to be highly correlated with Lp(a) even after adjustment for apo(a) isoforms. It still explained a relevant part (4%) of Lp(a) levels, especially for individuals with short isoforms, revealing a possible SNP-isoform interaction. The evaluation of correlation patterns within the LPA gene region and apo(a) isoforms and possible gene-gene and gene-environment interactions will also be a substantial part of this project. Finally, replicated SNPs will be validated on clinical outcomes (prospective myocardial infarctions) in the Copenhagen City Heart Study. Lp(a) has been shown to be a causal and independent risk factor for cardiovascular diseases in numerous studies. The proposed project would enable us to perform the most extended analyses on the genetic basis of Lp(a) done so far with the unique possibility to adjust and account for apo(a) isoforms. The major expectation for genome-wide association studies on Lp(a) is the identification of genes which are not yet known to influence Lp(a) levels and which might point to the largely unknown mechanism on the production, metabolism and catabolism of Lp(a). This might open new avenues for the therapeutic intervention of high levels of this highly atherogenic lipoprotein.

Lipoprotein(a) (Lp(a)) is an independent causal risk factor for cardiovascular diseases. Lp(a) is to a great extent genetically regulated by variants in the LPA gene cluster, a major part of it due to the K-IV-2 repeat, encoding apolipoprotein (a) (Apo(a)) isoforms. There is still a big gap between the estimated heritability and the variability which can be explained by already known genetic variants. To fill this gap, we performed the so far biggest genome-wide association study (GWAS) on Lp(a), based on data of 5 population-based studies including 13,000 participants. Since both Lp(a) and Apo(a) isoforms were measured in all participants, we were able to find specifically those genetic variants, which lead to Lp(a) values which cannot be explained by their Apo(a) isoforms. 40 single nucleotide polymorphisms (SNPs) in the LPA gene cluster und 1 SNP in the APOE gene were identified. 30 additional SNPs were associated with Lp(a) independent of Apo(a) isoforms. One of these variants emerged to be a marker for the effect of a hitherto unknown, but common, splice-site variant in the repetitive K-IV-2 region. This region is not accessible by common sequencing approaches. Therefore, this region, covering 70% of the LPA coding region, has been neglected in large parts so far. Therefore, we developed an ultra-deep sequencing approach to cover this region. We could newly describe more than 100 variants. The already mentioned splice-site variant was found to explain 20% of the Lp(a) variability in individuals with low-molecular weight Apo(a) (=small) isoforms and in consequence lowers the risk for cardiovascular diseases (CVD). For all identified SNPs, we could show a direct proportional relationship between their effect on Lp(a) and their CVD risk. The highest effect on Lp(a) was found for one rather rare SNP, leading to an increase of CVD risk by about 70% compared to those individuals not carrying this variant. Furthermore, using the statistical approach of so-called Mendelian randomization we were able to estimate the required reduction of Lp(a) by potential drugs to effectively reduce cardiovascular risk. To be equally effective as reducing LDL cholesterol by statins, Lp(a) would have to be lowered by 65 mg/dL. This has important implications for the planning of Lp(a) reducing medications.