This commentary refers to ‘Intestinal cholesterol and phytosterol absorption and the risk of coronary artery disease’ by J. Plat et al., 2021;42:281–282.
We appreciate the interest of Plat et al. in our paper.1 They argue that the coronary artery disease (CAD) risk associated with ABCG5/8 variants that is not accounted for by non-HDL cholesterol could be explained by several other lipid-related parameters including LDL sub-fractions and oxidized LDL. We note that genome-wide association studies (GWAS) of NMR-based measures2 , 3 demonstrate that ABCG5/8 variants have similar patterns of effects on LDL and VLDL sub-fractions as many non-HDL/LDL cholesterol associated variants in other genes including PCSK9, LDLR, and CELSR2. Differences in sub-fraction profiles of ABCG5/8 and other variants are therefore unlikely to account for the excess CAD risk. Also, a GWAS study on oxidized LDL levels4 did not show association with ABCG5/8 variants, while a missense variant in APOB associated with oxidized LDL, but not with cardiovascular events. Thus, the excess CAD risk for ABCG5/8 variants is also unlikely driven by their association with oxidized LDL. Plat et al. also argue that our study lacks consideration of other essential proteins in sterol transport. However, not investigating other proteins does not invalidate our results on genetic variation in ABCG5/8.
While our results do not prove a causal role of phytosterols in atherosclerotic disease, we consider elevated phytosterol levels driven through the ABCG5/8 variants the simplest and most plausible explanation for the excess CAD risk, for reasons thoroughly discussed in the manuscript. We do not agree that our conclusion and the translational perspective of our study needs revising.
Phytosterols have a log-normal distribution. In our data (N = 3039) (Table 1), the mean absolute stigmasterol, sitosterol, and campesterol levels were 0.033 [interquartile range (IQR) 0.020–0.066], 0.267 (IQR 0.172–0.414), and 1.618 (IQR 1.045–2.455) µg/mL, respectively (calculated on log-transformed values and transformed back to the original scale). One SD change in log-transformed phytosterol levels corresponds to 218%, 96%, and 96% change in absolute levels of stigmasterol, sitosterol, and campesterol, respectively. Thus, for example, the increasing sitosterol levels of 0.27 SD conferred by each allele of the intronic variant, corresponds to ∼26% change in absolute levels.
Table 1.
The mean absolute levels of phytosterols per genotype group (N = 3039)
| ABCG-[5/8] | Coding change | rsName[EA] | N per genotype (0/1/2) | Stigmasterol mean (µmol/mL) | Sitosterol mean (µmol/mL) | Campesterol mean (µmol/mL) |
|---|---|---|---|---|---|---|
| 5 | p.Phe624Leu | rs150401285[G] | 3020/19/0 | 0.033/0.095/- | 0.266/0.500/- | 1.611/2.981/- |
| 5 | p.His250Tyr | rs776502883[A] | 2989/50/0 | 0.033/0.102/- | 0.263/0.580/- | 1.601/2.946/- |
| 5 | p.Arg198Gln | rs141828689[T] | 2985/54/0 | 0.033/0.045/- | 0.266/0.315/- | 1.612/1.910/- |
| 5 | p.Gly27Ala | rs56204478[G] | 2988/51/0 | 0.033/0.068/- | 0.265/0.383/- | 1.609/2.248/- |
| 8 | p.Asp19His | rs11887534[C] | 2773/259/7 | 0.034/0.023/0.028 | 0.273/0.215/0.180 | 1.645/1.361/1.101 |
| 8 | Intronic | rs4299376[G] | 1520/1276/242 | 0.028/0.037/0.054 | 0.239/0.285/0.382 | 1.461/1.728/2.160 |
| 8 | p.Gln271Arg | rs770309304[G] | 2999/40/0 | 0.033/0.146/- | 0.263/0.834/- | 1.593/5.058/- |
| 8 | p.Arg263Gln | rs137852990[A] | 2967/70/2 | 0.032/0.093/3.879 | 0.264/0.420/12.650 | 1.599/2.566/13.333 |
| 8 | p.Trp361Ter | rs137852987[A] | 2980/59/0 | 0.033/0.065/- | 0.266/0.331/- | 1.609/2.116/- |
| 8 | p.Thr400Lys | rs4148217[A] | 1990/926/123 | 0.035/0.030/0.030 | 0.278/0.248/0.237 | 1.669/1.536/1.435 |
The mean absolute levels (calculated on log-transformed values and transformed back to the original scale) for non-carriers/heterozygous/homozygous carriers of the effect allele (EA).
When our study was performed, we did not measure cholesterol and cholestanol with the phytosterols. Cholesterol absorption biomarkers such as phytosterols or cholestanol provide information about whether elevated cholesterol levels are due to hyperabsorption, as opposed to hypersynthesis. In our study, the phytosterols are used for this purpose, so data on additional absorption biomarkers are not necessary. We also point out that cholestanol data would not inform about the ABCG5/8 associated CAD risk not explained by non-HDL cholesterol. In addition to the use of phytosterol levels as indicators of cholesterol absorption, previous studies support the atherogenic impact of these sterols themselves. The CAD risk that is not accounted for by non-HDL cholesterol for ABCG5/8 variants, may thus ‘relate to cholesterol absorption as such’, through their correlation with phytosterols.
Conflict of interest: none declared.
Contributor Information
Anna Helgadottir, deCODE Genetics/Amgen, Inc., Population Genomics, Sturlugata 8, Reykjavik, 102, Iceland; and.
Gudmar Thorleifsson, deCODE Genetics/Amgen, Inc., Population Genomics, Sturlugata 8, Reykjavik, 102, Iceland; and.
Kari Stefansson, deCODE Genetics/Amgen, Inc., Population Genomics, Sturlugata 8, Reykjavik, 102, Iceland; and; Faculty of Medicine, University of Iceland, Vatnsmyrarvegur, 101 Reykjavik, Iceland.
References
- 1.Plat J, Strandberg TE, Gylling H. Intestinal cholesterol and phytosterol absorption and the risk of coronary artery disease. Eur Heart J 2021;42:281–282. [DOI] [PubMed] [Google Scholar]
- 2. Chasman DI, Paré G, Mora S, Hopewell JC, Peloso G, Clarke R, Cupples LA, Hamsten A, Kathiresan S, Mälarstig A, Ordovas J, Ripatti S, Parker AN, Miletich JP, Ridker PM. Forty-three loci associated with plasma lipoprotein size, concentration, and cholesterol content in genome-wide analysis. PLoS Genet 2009;5:e1000730. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Tukiainen T, Kettunen J, Kangas AJ, Lyytikainen L-P, Soininen P, Sarin A-P, Tikkanen E, O'Reilly PF, Savolainen MJ, Kaski K, Pouta A, Jula A, Lehtimaki T, Kahonen M, Viikari J, Taskinen M-R, Jauhiainen M, Eriksson JG, Raitakari O, Salomaa V, Jarvelin M-R, Perola M, Palotie A, Ala-Korpela M, Ripatti S. Detailed metabolic and genetic characterization reveals new associations for 30 known lipid loci. Hum Mol Genet Hum Mol Genet 2012;21:1444–1455. [DOI] [PubMed] [Google Scholar]
- 4. Mäkelä K-M, Seppälä I, Hernesniemi JA, Lyytikäinen L-P, Oksala N, Kleber ME, Scharnagl H, Grammer TB, Baumert J, Thorand B, Jula A, Hutri-Kähönen N, Juonala M, Laitinen T, Laaksonen R, Karhunen PJ, Nikus KC, Nieminen T, Laurikka J, Kuukasjärvi P, Tarkka M, Viik J, Klopp N, Illig T, Kettunen J, Ahotupa M, Viikari JSA, Kähönen M, Raitakari OT, Karakas M, Koenig W, Boehm BO, Winkelmann BR, März W, Lehtimäki T. Genome-wide association study pinpoints a new functional apolipoprotein B variant influencing oxidized low-density lipoprotein levels but not cardiovascular events: atheroRemo consortium. Circ Cardiovasc Genet Circ Cardiovasc Genet 2013;6:73–81. [DOI] [PubMed] [Google Scholar]