Hegele, Robert A.
Western University, London, Ontario, Canada
Correspondence to Robert A. Hegele, MD, FRCPC, Robarts Research Institute, 4288A – 1151 Richmond Street North, London, ON N6A 5B7, Canada. Tel: +1 519 931 5271; fax: +1 519 931 5218; email: email@example.com
The first genome-wide association studies (GWASs) of plasma lipids emerged about 7 years ago . These initial efforts were carried out in relatively small samples of a few thousand individuals and were performed using early micorarrays that could genotype ∼100 000 single-nucleotide polymorphisms (SNPs). Although they yielded few significant findings, the first-generation GWASs signaled that much larger sample sizes were required in order to detect the very modest effects of genetic variants on plasma lipids, especially in epidemiological cohorts that were predominantly normolipidemic. The major conceptual and pragmatic advances that rapidly followed included: combining the results of many studies and samples, and meta-analysis of the associations across tens (or hundreds) of thousands of individuals; deployment of microarrays with more content, for example 1–2 million SNPs; and parallel development of bioinformatics and statistical methods, together with increased computational power and capacity to manage very large datasets.
The orchestration and execution of these extremely large GWAS projects over the subsequent years was a true tour de force. The amazingly consistent and replicable results not only confirmed the importance of well known loci and genes, but also identified significant new loci with comparable effects on plasma lipids. This culminated with the report in 2010 by the Global Lipids Genetics Consortium (GLGC) meta-analysis of 95 GWAS loci associated with plasma concentrations of LDL cholesterol (LDL-C), HDL cholesterol (HDL-C) and triglyceride or various combinations of these traits . Indeed, GWAS results in the lipid field were among the earliest and most impactful success stories supporting the overall approach.
But, although unquestionably significant, the observed effect sizes – that is the differences in mean lipid concentrations between individuals stratified by SNP genotypes – were quite modest. Cumulatively, these 95 loci explained ∼10% of the total variation in LDL-C, HDL-C and triglyceride levels in the general population . This explains why such large sample sizes were required. These same SNPs explained somewhat more – ∼25% – of genetic susceptibility in patients with pathological dyslipidemias, such as severe hypertriglyceridemia . But, although the GWAS results had clearly moved the lipoprotein field forward, a nagging question remained about the ‘missing heritability’, namely ‘Where in the genome does the remaining 75–90% of the variation reside?’ Proposed alternative sources of such heritability included: additional SNP loci; possible contributions of rare variants not represented on current microarrays, and which required either differently designed microarrays or DNA sequencing for their detection; a contributory role for gene-by-gene interactions or gene-by-environment interactions; and fundamentally different mechanisms altogether, such as mitochondrial genetic effects or epigenetics.
Now in 2014, a clearer although unpromising picture is emerging with respect to the ‘missing heritability’ for plasma lipids, at least that portion that is attributable to genomic sources. So far, it appears that the striking successes in finding new determinants seen with earlier GWASs have not quite been matched either by larger and more intense GWASs or by focusing on rare variants using next-generation sequencing (NGS).
The most recent update to the GWAS-generated list of genomic regions associated with plasma lipids was reported in 2013 by the updated GLGC version 2.0 , which has added 62 new loci, of which 30 were previously unknown, to the original 95 reported in GLGC version 1.0, giving a total of 157 loci . This was accomplished through the use of meta-analysis and extremely large datasets: up to 188 577 individuals, including large samples taken from multiple ethnicities. But despite studying almost twice as many individuals as GLGC version 1.0, there were not twice as many new loci identified in GLGC version 2.0. Furthermore, the effect sizes of the new loci were relatively modest: the new variants together explained an additional ∼2% of variation of the lipid traits. Many P values hovered around the nominal level of significance adjusted for multiple comparisons (P < 5 × 10–8). So the findings of GLGC version 2.0, although significant, indicate that the GWAS vein for lipids may be close to being tapped out. The field is still trying to grapple with the biological meaning of even just a few of the significant larger effect hits from GLGC version 1.0. And filling in the picture with additional loci, although important, has not solved the problem of the ‘missing heritability’.
Similarly, it appears that an answer will not soon be forthcoming from NGS to identify rare variants in patients or families with extreme lipoprotein phenotypes. By and large, the results so far of NGS studies have led to identification of causative mutations in genes that are already well known in monogenic dyslipidemias. The exception is ANGPTL3, in which compound heterozygous mutations cause familial combined hypolipidemia . But among individuals with severe but not classical familial hypercholesterolemia, NGS has so far identified causative mutations only in three known genes: ABCG5 encoding a sterolin half-transporter , APOE encoding apolipoprotein E [7,8] and LIPA encoding lysosomal acid lipase . Of course characterizing such minor genetic forms of hypercholesterolemia adds to our knowledge base, but it is still remarkable that after almost 3 years during which NGS has been a fairly standard research tool, and is increasingly used clinically, virtually no new disease genes have been identified in families with monogenic dyslipidemias.
Although the agnostic but hypothesis-generating GWAS and NGS approaches can expose both new and already known pathways and determinants, there remains a critical role for other types of focused experiments to add strands of understanding to the complex tapestry that is lipid metabolism. Several articles in this issue revisit and update current concepts of interesting – and important – entities that have been shown ‘the old-fashioned way’ – that is through biochemical, cellular and physiological experiments – to be mechanistically relevant.
Some of the most important recent advances have been in the area of the biology of adipocytes. In this regard, Kathrin Zierler, Rudi Zechner and Guenter Haemmerle (pp. 102–109) review current awareness of CGI-58 – an activator of adipose triglyceride lipase that regulates intracellular triglyceride levels and energy homeostasis. Carole Sztalryd synthesizes the emerging knowledge of perilipin 5, which regulates lipid droplet accumulation, promotes their association with mitochondria and also protects cells from fatty acid-associated toxicity. The theme of cytoprotection is echoed as Jin Ye and Hyeonwoo Kim scrutinize another new player – UBXD8 – that regulates lipogenic activity and acts as an intracellular sensor of unsaturated fatty acids, initiating reactions that thwart their overaccumulation. Moreover, Randy Kaufman (pp. 125–132) discusses new understandings of the importance of protein misfolding in the endoplasmic reticulum of hepatocytes and how accumulation of unfolded proteins leads to the unfolded protein response to protect the cell from stress. It turns out that a wide range of factors can modulate these intracellular pathways; some of these may be new therapeutic targets.
In the meantime, Andrew Brown, Elina Ikonen and Vesa Olkkonen (pp. 133–139) gauge the importance and possible mechanisms that explain how circulating cholesterol precursors – such as squalene, 7-dehydrocholesterol and desmosterol – might be early markers of cardiovascular disease and nonalcoholic steatohepatitis. Also, the emerging understanding of the linkages between bile acid metabolism, glucose metabolism, fatty liver and liver regeneration is reviewed by Karen Reue, Jessica Lee, and Laurent Vergnes (pp. 140–147), who focus particularly on the multifaceted roles of intestinal FGF15/19, and its relationship with Diet1, an intestinal factor that has a post transcriptional influence on FGF15/19 expression. Finally, Silvio Zaina (pp. 148–153) considers the evolving role of epigenetics in atherogenesis, focusing on DNA methylation, but also touching upon such newer entities as noncoding RNAs – both as mechanisms and potential markers of cardiovascular disease risk.
The material presented in this year's issue confirms the spectrum of interesting discovery and progress in the molecular biology and genetics of lipid metabolism. For the foreseeable future, I plan to closely monitor large-scale genotyping and sequencing efforts aimed at defining new genes and variants in physiological lipid metabolism, but also in abnormal clinical phenotypes and in rare familial disorders. Will upcoming studies find their second wind, reporting significant advances in the genetic determinants of plasma lipoproteins? Or will they mainly report refinements of genetic technologies, experimental designs or statistical analytical approaches? The latter are essential scientifically, especially to guide future endeavours in the genomics of less well studied disciplines. But if we are really reaching the limit of what ‘big genetics’ can teach us about the network of candidate players involved in lipid metabolism, then it would seem logical to rebalance some effort on more traditional biological approaches of function and mechanism. Perhaps it is time to evaluate in depth the wealth of results already in hand. The articles in this issue emphasize that physiological and functional experiments performed by smaller groups or networks of investigators working within well defined model systems can provide valuable contributions that ultimately move our field forward.
Conflicts of interest
There are no conflicts of interest.
1. Kathiresan S, Manning AK, Demissie S, et al. A genome-wide association study for blood lipid phenotypes in the Framingham Heart Study. BMC Med Genet. 2007; 8:(Suppl 1):S17
2. Teslovich TM, Musunuru K, Smith AV, et al. Biological, clinical and population relevance of 95 loci for blood lipids. Nature. 2010; 466:707–713.
3. Johansen CT, Wang J, Lanktree MB, et al. Excess of rare variants in genes identified by genome-wide association study of hypertriglyceridemia. Nat Genet. 2010; 42:684–687.
4. Global Lipids Genetics Consortium Discovery and refinement of loci associated with lipid levels. Nat Genet. 2013; 45:1274–1283.
5. Musunuru K, Pirruccello JP, Do R, et al. Exome sequencing, ANGPTL3 mutations, and familial combined hypolipidemia. N Engl J Med. 2010; 363:2220–2227.
6. Rios J, Stein E, Shendure J, et al. Identification by whole-genome resequencing of gene defect responsible for severe hypercholesterolemia. Hum Mol Genet. 2010; 19:4313–4318.
7. Marduel M, Ouguerram K, Serre V, et al. Description of a large family with autosomal dominant hypercholesterolemia associated with the APOE p.Leu167del mutation. Hum Mutat. 2013; 34:83–87.
8. Awan Z, Choi HY, Stitziel N, et al. APOE p.Leu167del mutation in familial hypercholesterolemia. Atherosclerosis. 2013; 231:218–222.
9. Stitziel NO, Fouchier SW, Sjouke B, et al. Exome sequencing and directed clinical phenotyping diagnose cholesterol ester storage disease presenting as autosomal recessive hypercholesterolemia. Arterioscler Thromb Vasc Biol. 2013; 33:2909–2914.