Background: Genome-wide association studies (GWAs) have identified almost one hundred genetic loci associated with variances in human blood lipid phenotypes including very low-density lipoprotein cholesterol, low-density lipoprotein cholesterol, high-density lipoprotein cholesterol, total cholesterol and triglycerides. Nevertheless the revealed loci only explain a small fraction of heritability and therefore a subtile phenotype of cholesterol homoestasis was examined in our study for the very first time. Methods and Results: Using a multi-stage approach of a GWA, firstly, a genome-wide analysis (Affymetrix 500K GeneChip) for serum lanosterol and serum total cholesterol using LC-MS/MS was conducted in 1495 participants of the KORA-S3/F3 cohort with subsequent replication in two additional independent samples of the the KORA-S3/F3 cohort (n = 1157) and CARLA cohort (n = 1760). Two genetic variants, SNP rs7703051 and rs17562686, in the HMGCR locus were significantly associated with serum lanosterol and showed similar effects of elevated serum lanosterol for each minor allele (combined n = 4412: p = 1,4 x 10-10, +7,1% and p = 4,3 x 10-6, +7,8%). Furthermore, rs7703051 showed a nominal statistical significance to serum cholesterol (p = 0,04). A combined analysis of both SNPs demonstrated that observed associations of rs17562686 can be partly explained by LD with rs7703051 being the primary polymorphism in that study. Nevertheless, rs17562686 shows consistent independent effects on serum lanosterol, thus being associated to a lipid phenotype for the very first time. The following SNP-fine mapping of the HMGCR locus was carried out in the CARLA cohort with subsequent validation in the LE-Heart cohort (n = 1895). The recently published SNP rs3846662 being in tight LD with rs7703051 could be associated with variances of serum lanosterol in both cohorts and functional in vivo studies of gen expression using qRT-PCR assays demonstrated a highly significant association of higher expression of alternatively spliced HMGCR mRNA lacking exon 13 with homozygosity for the rs3846662 major allele in 51 human liver samples (p < 0,01) and 958 human PBMCs (p = 2,1 x 10-7). The overall HMGCR gen expression was not affected. Further investigation of in vivo HMG-CoA reductase enzyme activity in both human samples (n = 48 and n = 55) using anionic exchange column chromatography and scintillation counting of [3-14C]-HMG-CoA and [5-3H]-mevalonolacton did not show any significant results. In addition there was not any association in the LE-Heart cohort between these SNPs and the development of CAD. Finally, rs7703051 could be replicated for already published total cholesterol (combined n = 4412) and rs3846662 for LDL-cholesterol (LE-Heart n = 1895). Since fine mapping in CARLA showed several SNPs throughout the HMGCR locus being in LD with rs17562686 we performed a DNA sequencing of the extended 5´-HMGCR promotor region in six human liver samples. A unknown SNP was discovered in the promotor but could not be associated with any of the examined phenotypes mentioned above. The minor allele of SNP rs5909 situated next to the stop codon and being in high LD with rs17562686 was associated with elevated serum lanosterol and slightly reduced HMGCR gen expression, but further studies including the above mentioned as well as measurement of 3’-UTR transcript lengths using qRT-PCR assays did not produce significant results. Conclusion: The phenotype serum lanosterol could be associated with genetic polymorphisms (e.g. rs7703051) in the HMGCR locus. Therefore already published associations of HMGCR with total cholesterol and LDL-cholesterol can be explained by variances of cholesterol homeostasis. The SNP rs17562686 could be associated with a phenotype of human blood lipids for the very first time. Subsequent gen expression analyses demonstrated a highly significant association of rs3846662 with variant patterns of HMGCR alternative splicing. A significant effect of alternatively spliced protein on enzyme activity and a association of these SNPs with CAD could not be shown.

Background: Genome-wide association studies (GWAs) have identified almost one hundred genetic loci associated with variances in human blood lipid phenotypes including very low-density lipoprotein cholesterol, low-density lipoprotein cholesterol, high-density lipoprotein cholesterol, total cholesterol and triglycerides. Nevertheless the revealed loci only explain a small fraction of heritability and therefore a subtile phenotype of cholesterol homoestasis was examined in our study for the very first time. Methods and Results: Using a multi-stage approach of a GWA, firstly, a genome-wide analysis (Affymetrix 500K GeneChip) for serum lanosterol and serum total cholesterol using LC-MS/MS was conducted in 1495 participants of the KORA-S3/F3 cohort with subsequent replication in two additional independent samples of the the KORA-S3/F3 cohort (n = 1157) and CARLA cohort (n = 1760). Two genetic variants, SNP rs7703051 and rs17562686, in the HMGCR locus were significantly associated with serum lanosterol and showed similar effects of elevated serum lanosterol for each minor allele (combined n = 4412: p = 1,4 x 10-10, +7,1% and p = 4,3 x 10-6, +7,8%). Furthermore, rs7703051 showed a nominal statistical significance to serum cholesterol (p = 0,04). A combined analysis of both SNPs demonstrated that observed associations of rs17562686 can be partly explained by LD with rs7703051 being the primary polymorphism in that study. Nevertheless, rs17562686 shows consistent independent effects on serum lanosterol, thus being associated to a lipid phenotype for the very first time. The following SNP-fine mapping of the HMGCR locus was carried out in the CARLA cohort with subsequent validation in the LE-Heart cohort (n = 1895). The recently published SNP rs3846662 being in tight LD with rs7703051 could be associated with variances of serum lanosterol in both cohorts and functional in vivo studies of gen expression using qRT-PCR assays demonstrated a highly significant association of higher expression of alternatively spliced HMGCR mRNA lacking exon 13 with homozygosity for the rs3846662 major allele in 51 human liver samples (p < 0,01) and 958 human PBMCs (p = 2,1 x 10-7). The overall HMGCR gen expression was not affected. Further investigation of in vivo HMG-CoA reductase enzyme activity in both human samples (n = 48 and n = 55) using anionic exchange column chromatography and scintillation counting of [3-14C]-HMG-CoA and [5-3H]-mevalonolacton did not show any significant results. In addition there was not any association in the LE-Heart cohort between these SNPs and the development of CAD. Finally, rs7703051 could be replicated for already published total cholesterol (combined n = 4412) and rs3846662 for LDL-cholesterol (LE-Heart n = 1895). Since fine mapping in CARLA showed several SNPs throughout the HMGCR locus being in LD with rs17562686 we performed a DNA sequencing of the extended 5´-HMGCR promotor region in six human liver samples. A unknown SNP was discovered in the promotor but could not be associated with any of the examined phenotypes mentioned above. The minor allele of SNP rs5909 situated next to the stop codon and being in high LD with rs17562686 was associated with elevated serum lanosterol and slightly reduced HMGCR gen expression, but further studies including the above mentioned as well as measurement of 3’-UTR transcript lengths using qRT-PCR assays did not produce significant results. Conclusion: The phenotype serum lanosterol could be associated with genetic polymorphisms (e.g. rs7703051) in the HMGCR locus. Therefore already published associations of HMGCR with total cholesterol and LDL-cholesterol can be explained by variances of cholesterol homeostasis. The SNP rs17562686 could be associated with a phenotype of human blood lipids for the very first time. Subsequent gen expression analyses demonstrated a highly significant association of rs3846662 with variant patterns of HMGCR alternative splicing. A significant effect of alternatively spliced protein on enzyme activity and a association of these SNPs with CAD could not be shown.