The study analysed the contribution of genetic variation to metabolomic measurements of plant sterols within a subset of 1,652 individuals for β-sitosterol and 1,451 individuals for campesterol measurements from a larger cohort of 6,218 Qatari participants. It is only the second such study worldwide, aimed at identifying genomic loci associated with sitosterolemia, a rare inherited disease, which can be misdiagnosed and/or masked by familial hypercholesterolemia.
ABCG5/8 are the major players across populations for both beta-sitosterol and campesterol.
Our results showed that 18 and 20 variants were significantly associated with β-sitosterol and campesterol traits, respectively, at the genome-wide significance level of P < 5.0×10− 8. Notably, our findings reconfirmed that the ABCG5/8 genes on the 2p21 locus were the primary predictors of sitosterol, with three out of four independent variants mapping within this region, consistent with previous research (34). Specifically, the results revealed that four independent SNPs, namely rs145164937 and rs4299376 for the β-sitosterol trait, and rs7598542, and rs75901165 for the campesterol trait, reached genome-wide significant associations (P-value < 5.0×10− 8), while two independent SNPs, namely rs72796720, rs781194036 had suggestive associations for β-sitosterol trait (P-value < 1.0×10− 5). All are novel SNPs except for rs4299376, which had been previously reported in the GWAS Catalog by Scholz et al. where the unconditional analysis associated it to the trait with a P-value of 9.5x10− 74 (23). Despite being mapped to intron 2 of ABCG8 with no clinical variant attributed in the ClinVar database, this SNP is one of the six independent variants derived from the European study that we used to devise our polygenic risk score. The allele frequency of allele 2 (T) of the variant in the QGP cohort is 0.553, whereas it was 0.680 in the Scholz study and 0.757 in the gnomAD. This highlights the slight differences between populations albeit the significant differences between the cohorts’ size. Functionally, the phenotypic measurements of plant sterol revealed that there are 774 individuals with the heterozygous variant, exhibiting an average β-sitosterol of 1.18, and 523 individuals with the homozygous variant, exhibiting an average β-sitosterol of 1.06. This was corroborated by eQTL analysis of the two alleles in human tissues, specifically in the colon whereby the TT genotype is correlated with higher expression of the transporter (Fig. 3).
The unique features related to the Qatari population were mostly highlighted by the identification of the other 3 novel SNPs in the same locus, in particular the rs145164937 variant. ClinVar shows conflicting interpretation of pathogenicity for this SNP. It is a novel finding as it has not been reported in the GWAS Catalog. Functionally, rs145164937 leads to an amino acid change in the N-terminal part of the protein (p.Gly98Ala) near the functional region implicated in binding ATP (Walker A). The QGP allele frequency for allele “C” (encoding Alanine) is 0.01, whereas its frequency amongst Europeans is 0.002 (Table S7). A significant P-value of 5.54×10− 14 indicated a strong association with the β-sitosterol levels Furthermore, the effect direction for the C allele (Beta = + 1.065), indicates that the presence of Alanine could lead to a decrease in the ABCG5 transporter function, and potentially increasing serum β-sitosterol levels. This reduced functionality was also reported in a large European cohort study including the UK biobank. The results showed that rs145164937 “C” allele was associated with increased risk of CAD (18).
The measurement of all plant sterols using metabolomics is crucial for diagnosing Sitosterolemia, but β-sitosterol represents the most important diagnostic criterion and is the predominant form of sterol found in Sitosterolemia patients (35). In this study, we aimed however to be more inclusive by adding campesterol in our genome-wide association analysis. Overall, our results provide new insights into the genetic architecture of sitosterol metabolism and may have implications for the diagnosis and treatment of sitosterolemia beyond the ABCG5/8 targeting strategies.
Expanding the genetic spectrum: SCARB1 and IFT43 have distinctive signatures.
While findings of the major role played by the ABCG5/8 were concordant for both campesterol and beta-sitosterol, small and significant divergence was found in two loci encoding the SCARB1 and IFT43 genes. Our study did consolidate the previous European findings on the novel role of SCARB1 by listing 3 variants (rs10846742, rs7485656, and rs112403212) as opposed to only one (rs10846744) by the previous study. While the functional role of the SCARB1 protein in sitosterol metabolism is not known, its role in reverse cholesterol transport is well documented highlighting the intimate relationship between plant and animal sterols in terms of absorption and excretion from the human body. The findings suggest that the addition of the new data provided stronger evidence for the association of SCARB1 with the trait. SCARB1 is a crucial component in the reverse cholesterol transport pathway and plays an essential role in lipid metabolism.
The rs75901165 variant on chr14-(76055310) maps to an intronic region in the Intraflagellar transport (IFT) gene. The IFT43 gene has been implicated in the regulation of cilium length and IFT (36). Mutations in IFT43 have been associated with ciliopathies, including skeletal (37) However, none of the previous studies involving human subjects had assessed associations with sterols. Our investigation yielded no association of this variant or any other variant within the IFT43 in the European cohort. Interestingly, a knock-out of this gene is embryonically lethal, but the heterozygous mice showed a statistically significant increase in circulating cholesterol level with a P-value of 1×10− 04 (Ift43 Mouse Gene Details | intraflagellar transport 43 | International Mouse Phenotyping Consortium (mousephenotype.org). Additional functional assays are needed to ascertain a role for IFT43 in sitosterolemia while a more extensive cohort is needed from the MENA region to confirm the genomics findings and cross-compared to other populations.
Whether the differences obtained between the two traits relate to the function of each sterol in terms of its chemical structure and mode of absorption and elimination is to be further studied using in vitro and in vivo mice models. Of note that in humans the hepatic clearance of campesterol is significant lower than that of beta-sitosterol (38), whereas it is stigmasterol accumulation and not the other two plant sterols that causes cardiac injury in mice (21)
Developing a PGS for early diagnosis
This study aimed to develop a PRS for sitosterolaemia. Due to the limited sample size, we were unable to construct the PRS based on identified independent variants and test it on the same data. Instead, we utilized the six independent sitosterol SNPs identified by Scholz et al. to calculate the PRS. Our findings (Fig. 4A) revealed a positive model fit of R2 = 0.085 (adjusted- R2 = 0.082, P = 1.97x10− 12) and significant correlation (r = 0.17, P = 6.3×10− 12) between β-sitosterol levels in the samples and the calculated PRS. It also revealed a positive model fit of R2 = 0.107 (adjusted- R2 = 0.102, P = 4.84×10− 9, Fig. 4B) and significant correlation (r = 0.15, P = 1.1×10− 9) between campesterol levels in the samples and the calculated PRS. While the significance of such scores is excellent, their application on other populations needs to be tested and other variants in specific populations like the IFT43 one in the Qatari population could add more accuracy to the results.
In conclusion, our study has identified four independent variants in the GWAS that are associated with sitosterolemia in the Qatari population. Our results provide new insights into the genetic architecture of sitosterol metabolism and may have implications for the diagnosis and treatment of sitosterolemia. Our findings serve as a foundational study for future comparative research in this area.