key: cord-0005229-i8fqjfww
authors: Guo, Yanhong; Garcia-Barrio, Minerva T.; Wang, Laiyuan; Chen, Y. Eugene
title: Experimental Biology for the Identification of Causal Pathways in Atherosclerosis
date: 2016-02-05
journal: Cardiovasc Drugs Ther
DOI: 10.1007/s10557-016-6644-7
sha: b965b106311249a730bf7192ed6d4a9caab46a1f
doc_id: 5229
cord_uid: i8fqjfww

More than 60 genomic loci have been implicated by genome-wide association studies (GWAS) and exome-wide association studies as conferring an increased risk of myocardial infarction and coronary artery disease (CAD). However, the causal gene and variant is often unclear. Using the functional analysis of genetic variants in experimental animal models, we anticipate understanding which candidate gene at a specific locus is associated with atherosclerosis and revealing the underlying molecular and cellular mechanisms, ultimately leading to the identification of causal pathways in atherosclerosis and may provide novel therapeutic targets for the treatment of atherosclerotic cardiovascular disease.

Atherosclerotic cardiovascular disease (AS-CVD) is the leading cause of morbidity and mortality in many industrialized countries. Multiple epidemiological studies have demonstrated genetic predisposition for atherosclerotic cardiovascular disease. In the Framingham study, a family history of coronary artery disease (CAD), stroke or peripheral arterial disease was associated with a 2.4-fold increased risk for CAD in man and 2.2-fold in women [1] . In individuals from families with premature CAD, heritability was estimated to be 92-100 %, whereas within families with older cases the heritability ranges from 15 % to 30 % [2] . Despite the recognition of the genetic predisposition to CAD for many years, recent advancements in the genomic tools and techniques including high throughput genome-wide sequencing are leading to the identification of previously unrecognized molecular pathways. The first large genome-wide association study (GWAS) for CAD was published in 2007 and identified the association of variants on chromosome 9p21 and CAD [3] [4] [5] .

In studies reported between 2007 and 2015, GWAS and exome-wide association studies have identified more than 60 significant variants strongly associated with CAD and MI, encoding more than 100 different genes [6] [7] [8] [9] [10] .

After several years deliberate yet gradual hypothesisdriven studies, GWAS and next generation sequencing have led to identification of genetic components of CAD and quantitative traits without relying on any prior biological hypothesis. In theory, this approach offers a great potential for the discovery of new causes and mechanisms of disease. Once a putative causal gene is identified by human genetic studies, overexpression and knockout studies in animal models can be used to further establish the biological function and underlying mechanisms of such gene. Table 1 shows 61 identified loci with genome-wide significance (P < 5 × 10 −8 ) by large-scale GWAS and exome-wide association studies. In the 61 CAD loci, only 15 loci were associated with the conventional risk factors and most loci harbor genes that have not yet been implicated in the pathogenesis of atherosclerosis. Identification of susceptibility variants and translational findings from genetic variants can lead to novel therapeutic treatment for CAD (Fig. 1 ). This [7] review will briefly discuss current biological approaches to identify the causal genes and pathways, functional studies to unravel the underlying mechanisms of disease, potential therapeutic targets to treat atherosclerotic cardiovascular disease.

It is widely recognized that high serum low-density lipoprotein cholesterol (LDL-C) levels play a critical role in the initiation and progression of atherosclerosis. Both Mendelian randomization studies and GWAS for CAD risks have indicated that LDL-C is a causal risk factor for AS-CVD [7, 11] . Evidence from clinical trials demonstrates that reducing circulating LDL-C levels by inhibition of 3-hydroxy-3-methyglutaryl coenzyme A (HMG-CoA) reductase, which is the rate-limiting key enzyme in the hepatic biosynthesis of cholesterol, can result in beneficial clinical outcomes in patients at risk of CAD. The LDL-C metabolism pathway is complex, and genetic defects in any of a number of steps in this pathway can lead to increased LDL-C levels.

PCSK9 First identified by Seidah and colleagues in 2003, the serine protease proprotein convertase subtilisin/kexin type 9 (PCSK9) has been associated with mild to severe hypercholesterolaemia and an increased risk of CAD [12] [13] [14] . Sanna et al. sequenced seven genes associated with LDL-C in 256 individuals with extreme LDL-C levels, confirmed in a sample of ∼10,000 Finnish and Norwegian individuals, and identified that rs11591147, which leads to a non-synonymous R46L change in exon 1 of PCSK9, is independently associated with LDL-C [15] . The mechanism by which PCSK9 gene defects cause autosomal dominant familial hypercholesterolemia (FH) is likely due to gain-of-function mutations. PCSK9 is a serine protease that binds to the epidermal growth factorlike repeat A on the LDL receptor. Upon binding to the LDL receptor, PCSK9 initiates internalization and lysosomal degradation of the LDLR/PCSK9 complex, resulting in decreased LDL receptors on the surface of hepatocytes and increased LDL-C concentrations in blood. Loss-of-function mutations lead to higher levels of the LDL receptor, lower LDL-C levels, and decreased cardiovascular risk [16] [17] [18] . This phenotype has been confirmed by gain-or loss-of-function studies in mice [19] [20] [21] . Overexpression of PCSK9 diminishes LDL receptor protein and function, causing an Ldlr knockout phenotype in mice [19, 20] . Pcsk9-deficient mice showed reduced plasma cholesterol levels by increasing LDL receptor protein expression in liver and accelerating the clearance of circulating cholesterol [19, 21] . In apolipoprotein E (Apoe)-deficient mice, gene inactivation of Pcsk9 significantly reduced aortic cholesteryl esters. By comparison, overexpression of PCSK9 induced an accumulation of aortic cholesteryl esters and accelerated development of atherosclerotic plaque [19] . In Ldlrdeficient mice, overexpression of PCSK9 did not regulate the circulating cholesterol levels, aortic accumulation of cholesteryl esters and atherosclerotic plaque size, indicating that the pro-atherosclerotic effect of PCSK9 is mediated mainly through its action on the LDL receptor [19] . The PCSK9 transgenic pigs created by transposition of a human PCSK9 D374Y gain-of-function mutation displayed reduced hepatic LDL receptor levels, impaired LDL clearance, severe hypercholesterolemia, and spontaneous development of progressive atherosclerotic lesions [22] . Although PCSK9 gain-of-function mutations are rare, comprising only a small portion (up to 3 %) of cases of FH, the established mechanism has generated significant interest in the gene as a drug target. Both alirocumab and evolocumab, humanized monoclonal antibodies causing inactivation of PCSK9, result in large reductions in LDL cholesterol levels, as compared with placebo (39 to 62 % reduction for alirocumab and 47 to 56 % for evolocumab) [23, 24] . Although the potential cardiovascular benefits of these agents need to be further established, this newly confirmed pathway identified by human genetics has become a novel therapeutic target to lower cholesterol in patients with familial hypercholesterolemia and clinical atherosclerotic cardiovascular disease.

SORT1 In addition to Mendelian randomization studies, GWA studies have identified novel loci associated with CAD and LDL-C. One of the most compelling findings is a locus on chromosome 1p13, which is strongly associated with both plasma LDL-C and myocardial infarction (MI) in humans [25] [26] [27] . This region harbors seven genes, seryl tRNA synthetase (SARS), cadherin EGF lag seven-pass G type receptor 2 (CELSR2), proline serine rich coiled coil protein 1 (PSRC1), myosin binding protein H-like (MYBPHL), sortilin-1 (SORT1), proteasome (prosome, macropain) subunit, alpha type 5 (PSMA5), and synaptophysin-like 2 (SYPL2). Using expression quantitative trait loci (eQT) data, several groups confirmed that the minor allele for the SNP rs646776 and the major allele for SNP rs599839, were associated with low mRNA levels for SORT1, PSRC1 and CELSR2 in the liver [25, 28, 29] . Celsr2 is highly expressed in cells lining all cerebral ventricles and the central canal of the spinal cord [30] . Results from Celsr2-deficient mice indicated impaired cilia function and hydrocephalus, but there was not informative data regarding any effects in cholesterol level [30] . Musunuru et al. demonstrated that Sort1 overexpression led to a 40 % reduction in plasma cholesterol, whereas Psrc1 overexpression had no effect on plasma cholesterol using adeno-associated viruses (AAVs)-mediated overexpression of Psrc1 and Sort1 in hyperlipidemic Ldlr-deficient mice [29] . Sort1 knockdown in mouse liver using siRNAs was associated with a 20-40 % increase in plasma cholesterol [29] . These results strongly implicated SORT1 as the causal gene at the 1p13 locus whose elevated expression confers reduction in LDL-C. However, Kjolby et al. reported different results using adenovirus-mediated overexpression of SORT1 and Sort1-deficient mice, in which Sort1-deficient mice show decreased cholesterol levels and reduced atherosclerotic plaque area [31] . Recently, Mortensen et al. demonstrated that transfer of Sort1-deficient bone marrow into irradiated atherosclerotic mice reduced atherosclerosis and systemic markers of inflammation without concomitant changes in total cholesterol or LDL-C, indicating SORT1 plays an additional inflammatory role in lymphocytes and macrophages potentially contributing to overall atherogenesis [32] . Although the SORT1 locus represents a promising target for the treatment of atherosclerosis identified by human genetic studies, these obviously conflicting findings in animals and the yet unsettled mechanisms behind the phenotype rises some questions on their immediate translation into the therapeutic strategies.

TM6SF2 Both GWA studies and exome-wide association studies have identified that a nonsynonymous variant (rs58542926, encoding p.Glu167Lys) in transmembrane 6 superfamily member 2 (TM6SF2) gene is associated with blood lipid levels, nonalcoholic fatty liver disease (NAFLD) and MI [33] [34] [35] [34] . In mice, transient overexpression of human wildtype TM6SF2 resulted in high total cholesterol, LDL-C and TG, whereas knockdown of mouse Tm6sf2 resulted in decreased serum total cholesterol levels [34] . Our results demonstrate that TM6SF2 is a critical regulator of cholesterol metabolism and liver-specific Tm6sf2 knockout mice challenged to high fat diet did not show increased liver triglyceride content [36] . Recently, a systematic meta-analysis in 101,326 individuals showed that carriers of the minor T allele are protected from cardiovascular disease, showing lower levels of total cholesterol, LDL-C and triglyceride. In contrast, carriers of the T allele showed a moderate increased risk for NAFLD [37] . Based on the opposing roles of TM6SF2 on both CVD and NAFLD, it is necessary to dissect the two functions of TM6SF2 before it becomes a therapeutic target to reduce the risk of myocardial infarction via lowering blood lipids.

The existence of a strong inverse relationship between plasma HDL and CAD has been largely confirmed in numerous epidemiological studies. HDLs display a wide spectrum of activates including reverse cholesterol transport, anti-inflammatory and anti-oxidative effects and maintenance of endothelial function. Although most of the HDL-raising agents, such as niacin, fibrates, and cholesteryl ester transfer protein (CETP) inhibitors, have yielded convincing results to increase HDL-C levels, those clinical trials have failed to show a decrease in CAD events. Further evidence from genetic studies has largely failed to support a strong causal association between genetically raised plasma HDL-C levels and CAD risks using HDL-associated variants at cholesteryl ester transfer protein (CETP), lecithin-cholesterol acyltransferase (LCAT), ATP-binding cassette transporter A1 (ABCA1), and hepatic triglyceride lipase (LIPC). It has been questioned whether the association between HDL-C levels and the risk of atherosclerotic events represents a causal relationship. GWA studies and Mendelian randomization studies have suggested that genetic mechanisms that raise plasma HDL-C do not seem to lower risk of MI [38] . However, based on family and twin studies, HDL-C levels have shown a heritability ranging from 40 to 60 % [39, 40] . Recent GWA studies have shown that only 10-12 % of the heritability is due to common variants [41, 42] , suggesting that a large proportion of the molecular mechanisms underlying HDL-C functions and metabolism is still unknown.

KLF14 More than 10 HDL-associated loci with significant effect on CAD risk have been identified-CETP, LPL, LIPG, LPL, TRIB, APOA5-APOA4-APOC3-APOA1 cluster, PPP1R3B, TTC39B, GALNT2, KLF14, IRS, ANKS1A, and SH2B3. Among these genes only KLF14 and ANKS1A were solely associated with HDL-C levels and CAD. Teslovich et al. reported that variants at the locus containing KLF14 and TSGA13, which encode Krüppel-like factor 14 and testis-specific gene A13, were associated with both HDL-C and CAD [41] . These variants were maternally restricted cis-eQTLs for KLF14 expression and strongly correlated with expression of several other genes in adipose tissue in trans, suggesting that KLF14 may be a master regulator of gene expression in adipose tissue and a key player in human metabolism [43] . Guo et al. demonstrated that adenovirusmediated overexpression of KLF14 markedly increased HDL-C levels and cholesterol efflux capacity through regulation of hepatic ApoA-I production. Studies of Klf14 knockout mouse models have confirmed the relationship between Klf14 and HDL-C levels [44] . Furthermore, Guo and colleagues identified perhexiline, an approved therapeutic small molecule presently in clinical use to treat angina and heart failure, as a KLF14 activator. Treatment with perhexiline increased HDL-C levels and cholesterol efflux capacity and reduced atherosclerotic lesion development in Apoe-deficient mice [44] . This finding was directionally consistent with the human genetics data showing that the T allele of rs4731702 located near the KLF14 gene was associated with a decreased risk of myocardial infarction and the T allele carriers have higher ApoA-I levels in the Mulao population in China [45] . These data were crucial in establishing KLF14 as a causal contributor to HDL-C levels and risk of CAD and led to a rapid effort to understand its function and the therapeutic potential of targeting this protein.

For decades, the independent association between increased plasma triglyceride levels and cardiovascular risk had been largely controversial. Several randomized controlled clinical trials targeting triglyceride lowering such as fibrates failed to show beneficial effects on cardiovascular events. However, growing evidence from human genetic studies has shown a strong association between triglyceride and CAD risk. Do and colleagues analyzed 185 SNPs in 188,577 individuals from 60 studies and found that SNPs' effect on triglyceride levels is associated with risk of CAD, even after accounting for effects on LDL-C and/or HDL-C levels [46] . Furthermore, a multiple SNP Mendelian randomization analysis in over 62,000 participants (including 12,000 CAD events) demonstrated a causal relationship between triglycerides and CAD events [11] .

A locus on chromosome 11 is significantly associated with triglyceride, LDL-C, HDL-C and CAD risk identified by GWA studies. This complex cluster includes the apolipoprotein genes APOA1, APOC3, APOA4, and APOA5. ApoC-III inhibits lipoprotein lipase, promotes VLDL production and slows hepatic VLDL uptake [47, 48] . The Exome Sequencing Project performed an exome sequencing of 18, 666 genes in 3734 individuals and conducted an association assay for the rare variants with triglyceride levels [49] . They identified three loss-of-function mutations in APOC3. The mutation carriers have 39 % lower in triglyceride levels and 40 % reduced risk of CAD than noncarriers [49] . Similar findings were confirmed by resequencing the coding regions and consensus splice sites of APOC3 in the general population in Denmark [50] . Although the effect of ApoC-III on VLDL production in mice is somewhat controversial [51] [52] [53] [54] , antisense inhibition of apoC3 has been developed and administration of the apoC3 antisense oligonucleotide ISIS 304801 significantly decreased levels of apoC3 and triglycerides [55, 56] . ISIS 304801 treatments also resulted in increased HDL-C levels and reduced VLDL levels. Long-term outcome clinical trials targeted at lowering ApoC-III are required to establish the causality and to evaluate whether ApoC-III may serve as a CAD therapeutic target.

APOA5 GWA studies reveal the importance of ApoA-V as a regulator of plasma triglyceride homeostasis. ApoA-V accelerates lipoprotein lipase (LPL)-mediated hydrolysis of triglyceride-rich VLDL, inhibits VLDL production and facilitates VLDL clearance [57, 58] . APOA5 transgenic mice have significantly reduced serum triglyceride levels and Apoa5-deficient mice have 18-fold higher VLDL levels compared to control mice [59] . Recently, an exosome sequencing study with 6721 subjects with early-onset myocardial infarction and controls identified that carriers of rare non-synonymous mutations on APOA5 had higher plasma triglyceride levels and had a 2.2-fold increased risk for MI [60] . Those findings indicate that apoA5 may become a promising therapeutic target for triglyceridemia and CAD.

TRIB1 Lead SNPs located at chromosome 8p24 are associated with all major lipid traits and CAD in both GWAS and gene-centric array association studies, in which the minor allele of rs2954029 is associated with reduced triglycerides, decreased LDL-C, increased HDL-C, and reduced risk of CAD [41] . Adenovirus-mediated hepatic overexpression of Trib1 resulted in reduced plasma cholesterol and triglyceride levels, accompanied by reduced expression of key genes involved in lipogenesis, such as fatty acid synthase, stearoylcoenzyme A desaturase 1 and acetyl-CoA carboxylase 1. In contrast, Trib1 deficient mice showed increased plasma cholesterol and triglyceride levels [61] . Recently, carbohydrateresponsive element-binding protein (ChREBP) and Sin3A (Swi-independent 3 A)-associated protein emerged as TRIB1 binding partners [62, 63] , which may aid discovery of novel therapeutic targets for the management of dyslipidemia. Not known animal experimental results are available to illustrate whether TRIB1 is a causal gene for atherosclerosis at this time.

In the initial Ottawa population of 322 cases versus 312 controls, a locus on chromosome 9p21 was identified to be associated with CAD [3] [4] [5] . This association with CAD was confirmed in diverse ethnic groups except for Africans. The 9p21 genetic risk variant is extremely common, with 75 % of humans having one or more risk alleles (50 % heterozygous and 25 % homozygous). The presence of a single copy or two copies of the 9p21 risk allele conferred a 15-20 % or 30-40 % increased risk for CAD respectively. In individuals with premature CAD, 9p21 homozygosis is associated with a 2-fold increased risk for CAD [4] . The 9p21 risk is independent of conventional risk factors such as cholesterol, diabetes, or hypertension, indicating that 9p21 exerts its effect via some novel unknown pathway, potentially leading to new therapeutic targets. However, the mechanism whereby the 9p21 locus confers increased CAD risk remains obscure. Whole 9p21 region deletion mice confirmed that the 9p21 region is a regulator of the Cyclin-dependent kinase inhibitors, but the mice do not develop atherosclerosis [64, 65] . Fine mapping of the genetic risk locus at 9p21 shows that it consists of a cluster of 59 linked SNPs over a 53 kb region. The nearest protein coding genes are the cyclin-dependent kinase inhibitors CDKN1A, CDKN2B and methylthioadenosine phosphorylase (MTAP), which are abundantly expressed in atherosclerotic tissues and regulate cell cycle and apoptosis [66] . Deletion of the same region in the murine genome was associated with severe reduced expression of Cdkn2a and Cdkn2b mRNAs. In experimental atherosclerotic models, Cdkn2a-deficient mice have smaller lesions and loss of Cdkn2b is associated with advanced atherosclerotic lesions [67, 68] . Mtap-knockdown mice also have a significant increase in atherosclerotic lesions [67] . Another proposed underlying mechanism rests in a long noncoding RNA in the INK4 locus (ANRIL), which overlaps CDKN2B. Unlike CDKN1A and CDKN2B, ANRIL expression is highly associated with the 9p21 haplotype and decreased ANRIL is directly correlated with severity of common carotid artery stenosis [69] . In vitro studies have demonstrated that ANRIL regulates CDKN2B expression in cultured human fibroblasts. A significant limitation of the mouse atherosclerosis models to investigate the biological functions of Anril is the fact that mice lack a clear ortholog of ANRIL [70] . Nonhuman primates may be needed to illustrate the susceptibility to cardiovascular disease at the 9p21 locus. Recently, Harismendy and colleagues have identified 33 enhancers in 9p21 locus and two SNPs within the 9p21 locus which disrupt the binding for signal transducer and activator of transcription 1 (STAT1) to the enhancer that physically interacts with the CDKN2A/B locus and MTAP gene [73] . As STAT1 is a downstream inflammatory effector, these findings indicate that the 9p21 locus may play a role in inflammatory signaling in the arterial wall. Ongoing fundamentally translational research on the 9p21 locus will reveal the new underlying mechanisms and may drive development of novel therapeutic approaches for atherosclerosis.

ADAMTS-7, a member of the disintegrin and metalloproteinase with thrombospondin motifs (ADAMTS) family, was recently identified to be significantly associated with human CAD, restenosis or arterial calcification [9, 10, 71, 72] . As an enzyme, ADAMTS-7 affects the extracellular matrix of vascular smooth muscle cells, but this locus has no relationship to traditional risk factors. Overexpression of ADAMTS7 induced vascular smooth muscle cells migration and promoted neointima formation following artery injury through the degradation of cartilage oligomeric matrix protein (COMP) in rats [73] . Kessler and colleagues demonstrated that Adamts7 inhibits endothelial cell proliferation and migration via degradation of thrombospondin-1 in Comp-independent manner [74] . Adamts7-deficient mice showed enhanced vessel re-endothelialization and reduced neointima formation after wire injury [74] . Bauer and colleagues addressed the effect of Adamts7 on atherosclerosis development. Deletion of Adamts7 significantly reduced atherosclerotic lesion formation in mice on both Apoe knockout and Ldlr knockout background without significant changes in plasma lipid profiles [75] . Adamts7 deletion impaired VSMC migration and developed plaques with a larger fibrous cap, indicating targeting ADAMTS7 may result not only in decreased atherosclerosis but also in more stable plaques.

Over the last few decades, it is well accepted that inflammatory processes not only promote initiation of atherosclerosis, but also contribute to acute cardiovascular events, such as triggering MI. Network analysis demonstrates that some genes, such as APOA-I and Interleukin-6 receptor (IL-6R), involved in lipid metabolism pathways also involved in inflammation, contributing to CAD pathogenesis [9] . Increased Interleukin-6 (IL-6) expression has been associated with CAD. IL-6 binds to its receptor and plays a central role in propagating the downstream inflammatory response during atherosclerosis progression. A non-synonymous IL-6R variant is associated with increased circulating levels of IL-6 and decreased CAD risk [76, 77] , suggesting a causal relationship between IL-6R-related pro-inflammatory pathways and CAD. Tocilizumab, a monoclonal antibody that blocks both membrane-bound and circulating IL6R, has antiinflammatory actions and is licensed for treatment of rheumatoid arthritis. However, rheumatoid arthritis patients treated with an IL-6R monoclonal antibody, tocilizumab, show increased circulating triglycerides and LDL-C levels [78] , suggesting that blockage of IL-6R pathway to protect against atherosclerosis need to be reconsidered. The chemokine CXCL12 and its receptor, CXCR4, play an important protective role in atherosclerosis. Circulating levels of CXCL12 are decreased in CAD patients compared with healthy control subjects [79] . Rs1746048, at chromosome 10q11 near the CXCL12 gene, is associated with CAD and plasma levels of CXCL12 [80] . The CXCL12/CXCR4 axis is critical for retention and release of hematopoietic cells from the bone marrow and Cxcr4 deficient mice showed increased atherosclerotic lesions due to hyperactivation of circulating neutrophils [81, 82] . Other GWAS-identified signal for CAD includes ABO locus on chromosome 9q34 and the major histocompatibility complex (MHC) locus on chromosome 6p21.3 [71, 83, 84] . Further studies are required to confirm the association and to identify the causative allele.

The NO-sGC-cGMP pathway is a major pathway controlling vascular smooth-muscle relaxation, vascular tone, and vascular remodeling. Under normal conditions, NO binds to soluble guanylate cyclase (sGC) and activates sGC to generate cGMP, and thus stimulates cGMP-dependent protein kinase I, which inhibits VSMC constriction, proliferation and migration, and reduces platelet adhesion and activation as well as vascular inflammation [85] . Lu et al. performed a two-stage GWAS of CAD in a sample including 33,000 Han Chinese individuals, and identified that rs1842896 at 4q32.1 is associated with CAD [86] . GUCY1A3, encoding α1-sGC, is the only gene in this CAD-associated locus. Erdmann et al. reported the segregation of two private, heterozygous mutations in GUCY1A3 and CCT7, in an extended myocardial infarction family [87] . CCT7 encodes CCTη, a member of the tailless complex polypeptide 1 ring complex, which stabilizes α1-sGC. In vitro studies demonstrated that mutations in both GUCY1A3 and CCT7 severely reduce α1-sGC as well as β1-sGC protein content, and impaired α1-sGC activity. Moreover, platelets from mutation carriers contained less α1-sGC protein and consequently displayed reduced nitric-oxide-induced cGMP formation. Mice deficient in α1-sGC protein displayed accelerated thrombus formation in the microcirculation after local trauma [87] . These genetic and functional studies identified a link between impaired α1-sGC-dependent NO signaling and MI risk, possibly through accelerated thrombus formation. Studies are ongoing to reveal the role of Gucy1a3 in atherosclerosis as well as lipid metabolism.

GWA studies, exome sequencing, linkage analyses, and targeted resequencing have successfully identified potential causal variants. However, the understanding of causality is still rudimentary and the task of validation of causal genes and determining causal variants remains challenging. The most widely used in vivo model for atherosclerosis study is currently the mouse with Ldlr knockout or Apoe knockouts at the forefront. However, most of the GWAS identified susceptibility variants are found within noncoding regions, making it difficult to define disease associations with specific transcripts. In addition, the regulatory variants may not be evolutionarily conserved, for example, the mouse's lack of Cetp and a clear ortholog of ANRIL in the 9p21 region. Some large animal models, such as pigs, rabbits, and nonhuman primates, have particular advantages for certain types of studies because they have shown an atherosclerotic pattern comparable to that of humans [88] . Others and our group have generated transgenic rabbits and pigs with altered expression of human genes and specific gene knockout rabbits and pigs for cardiovascular disease studies [22, [89] [90] [91] [92] .

Several groups have shown that induced pluripotent stem cells (iPS cells) can be generated from human monocytes by the introduction of defined sets of transcription factors. These iPS cells resemble embryonic stem cells in many respects and will provide a powerful in vitro model for the study of gene function. Rader's group used iPS cell-derived hepatocytes from the SR-BI P376L mutation homozygous subject to study the effects of the endogenous mutation in human hepatocytelike cells (AHA2015 presentation). Engineered gene editing such as zinc finger nucleases and CRISPR-Cas9 systems can facilitate precise modification of the genome in iPS cells to obtain potential causal DNA variants. Directed differentiation of iPS cells makes it possible to establish cellular models with specific variants and to investigate relationships between the variants and specific metabolic pathways in appropriate human cell types.

In the past decade, GWA studies, exome sequencing, and targeted resequencing have offered a rapid increase in the number of CAD susceptibility loci. The truly established causal associations between genetic variation and CAD, such as PCSK9, have been becoming novel therapeutic targets. However, the mechanisms underlying the association of most of these loci to CAD remain unknown. By combination data from genetic variants, functional studies, genetic animal models, and drug development, ongoing studies are likely to accelerate our understanding of mechanisms linking identified loci to atherosclerosis, facilitate personalized medicine and lead to the development of novel therapeutic strategies for prevention and treatment of CAD.

Familial lipoprotein disorders in patients with premature coronary artery disease

Familial occurrence of coronary heart disease: effect of age at diagnosis

A common allele on chromosome 9 associated with coronary heart disease

A common variant on chromosome 9p21 affects the risk of myocardial infarction

p15Ink4b is a critical tumour suppressor in the absence of p16Ink4a

The 9p21 susceptibility locus for coronary artery disease and the severity of coronary atherosclerosis

A comprehensive 1000 genomes-based genome-wide association meta-analysis of coronary artery disease

Identifying novel gene variants in coronary artery disease and shared genes with several cardiovascular risk factors

Large-scale association analysis identifies new risk loci for coronary artery disease

Coronary Artery Disease Genetics C. A genome-wide association study in Europeans and south Asians identifies five new loci for coronary artery disease

Mendelian randomization of blood lipids for coronary heart disease

The secretory proprotein convertase neural apoptosis-regulated convertase 1 (NARC-1): liver regeneration and neuronal differentiation

Mutations in PCSK9 cause autosomal dominant hypercholesterolemia

The influence of PCSK9 polymorphisms on serum low-density lipoprotein cholesterol and risk of atherosclerosis

Fine mapping of five loci associated with low-density lipoprotein cholesterol detects variants that double the explained heritability

Missense mutations in the PCSK9 gene are associated with hypocholesterolemia and possibly increased response to statin therapy

Sequence variations in PCSK9, low LDL, and protection against coronary heart disease

Molecular characterization of loss-of-function mutations in PCSK9 and identification of a compound heterozygote

Gene inactivation of proprotein convertase subtilisin/ kexin type 9 reduces atherosclerosis in mice

Adenoviral-mediated expression of Pcsk9 in mice results in a low-density lipoprotein receptor knockout phenotype

Decreased plasma cholesterol and hypersensitivity to statins in mice lacking Pcsk9

Familial hypercholesterolemia and atherosclerosis in cloned minipigs created by DNA transposition of a human PCSK9 gain-of-function mutant. Science translational medicine

Efficacy and safety of alirocumab in reducing lipids and cardiovascular events

Efficacy and safety of evolocumab in reducing lipids and cardiovascular events

Six new loci associated with blood low-density lipoprotein cholesterol, high-density lipoprotein cholesterol or triglycerides in humans

Newly identified loci that influence lipid concentrations and risk of coronary artery disease

Common variants at 30 loci contribute to polygenic dyslipidemia

Mapping the genetic architecture of gene expression in human liver

From noncoding variant to phenotype via SORT1 at the 1p13 cholesterol locus

Lack of cadherins Celsr2 and Celsr3 impairs ependymal ciliogenesis, leading to fatal hydrocephalus

Sort1, encoded by the cardiovascular risk locus 1p13.3, is a regulator of hepatic lipoprotein export

Targeting sortilin in immune cells reduces proinflammatory cytokines and atherosclerosis

Exome-wide association study identifies a TM6SF2 variant that confers susceptibility to nonalcoholic fatty liver disease

Systematic evaluation of coding variation identifies a candidate causal variant in TM6SF2 influencing total cholesterol and myocardial infarction risk

TM6SF2 is a regulator of liver fat metabolism influencing triglyceride secretion and hepatic lipid droplet content

Hepatic transmembrane 6 superfamily member 2 regulates cholesterol metabolism in mice

A genome-wide association study confirms PNPLA3 and identifies TM6SF2 and MBOAT7 as risk loci for alcohol-related cirrhosis

Plasma HDL cholesterol and risk of myocardial infarction: a mendelian randomisation study

Effects of household sharing on high density lipoprotein and its subfractions

Genetic heritability and common environmental components of resting and stressed blood pressures, lipids, and body mass index in Utah pedigrees and twins

Biological, clinical and population relevance of 95 loci for blood lipids

Large-scale gene-centric meta-analysis across 32 studies identifies multiple lipid loci

Identification of an imprinted master trans regulator at the KLF14 locus related to multiple metabolic phenotypes

Perhexiline activates KLF14 and reduces atherosclerosis by modulating ApoA-I production

Association of the KLF14 rs4731702 SNP and serum lipid levels in the Guangxi mulao and Han populations

Common variants associated with plasma triglycerides and risk for coronary artery disease

Expression of apolipoprotein C-III in McA-RH7777 cells enhances VLDL assembly and secretion under lipid-rich conditions

Apolipoprotein C-III deficiency accelerates triglyceride hydrolysis by lipoprotein lipase in wild-type and apoE knockout mice

Loss-of-function mutations in APOC3, triglycerides, and coronary disease

Loss-of-function mutations in APOC3 and risk of ischemic vascular disease

Overexpression of human apolipoprotein C-III in transgenic mice results in an accumulation of apolipoprotein B48 remnants that is corrected by excess apolipoprotein E

Chylomicronemia due to apolipoprotein CIII overexpression in apolipoprotein E-null mice. Apolipoprotein CIII-induced hypertriglyceridemia is not mediated by effects on apolipoprotein E

Willems van dijk K. ApoC-III deficiency prevents hyperlipidemia induced by apoE overexpression

Apoprotein C-III deficiency markedly stimulates triglyceride secretion in vivo: comparison with apoprotein E

Antisense inhibition of apolipoprotein C-III in patients with hypertriglyceridemia

Antisense oligonucleotide inhibition of apolipoprotein C-III reduces plasma triglycerides in rodents, nonhuman primates, and humans

Apolipoprotein AV accelerates plasma hydrolysis of triglyceride-rich lipoproteins by interaction with proteoglycanbound lipoprotein lipase

ApoAV reduces plasma triglycerides by inhibiting very low density lipoprotein-triglyceride (VLDL-TG) production and stimulating lipoprotein lipase-mediated VLDL-TG hydrolysis

Effect of apolipoprotein A-V on plasma triglyceride, lipoprotein size, and composition in genetically engineered mice

Exome sequencing identifies rare LDLR and APOA5 alleles conferring risk for myocardial infarction

Trib1 is a lipid-and myocardial infarction-associated gene that regulates hepatic lipogenesis and VLDL production in mice

Tribbles-1 regulates hepatic lipogenesis through posttranscriptional regulation of C/EBPalpha

Sin3A-associated protein, 18 kDa, a novel binding partner of TRIB1, regulates MTTP expression

Functional analysis of the chromosome 9p21.3 coronary artery disease risk locus

Targeted deletion of the 9p21 non-coding coronary artery disease risk interval in mice

Expression of Chr9p21 genes CDKN2B (p15(INK4b)), CDKN2A (p16(INK4a), p14(ARF)) and MTAP in human atherosclerotic plaque

Effect of 9p21.3 coronary artery disease locus neighboring genes on atherosclerosis in mice

Cyclin-dependent kinase inhibitor 2B regulates efferocytosis and atherosclerosis

Genetic variants at the 9p21 locus contribute to atherosclerosis through modulation of ANRIL and CDKN2A/ B

Quest for genes and mechanisms linking the human chromosome 9p21.3 locus to cardiovascular disease

Identification of ADAMTS7 as a novel locus for coronary atherosclerosis and association of ABO with myocardial infarction in the presence of coronary atherosclerosis: two genome-wide association studies

Large-scale association analysis identifies 13 new susceptibility loci for coronary artery disease

ADAMTS-7 mediates vascular smooth muscle cell migration and neointima formation in balloon-injured rat arteries

ADAMTS-7 inhibits re-endothelialization of injured arteries and promotes vascular remodeling through cleavage of thrombospondin-1

Knockout of Adamts7, a novel coronary artery disease locus in humans, reduces atherosclerosis in mice

Receptor Mendelian Randomisation Analysis C. The interleukin-6 receptor as a target for prevention of coronary heart disease: a mendelian randomisation analysis

Interleukin-6 receptor pathways in coronary heart disease: a collaborative meta-analysis of 82 studies

Effect of interleukin-6 receptor inhibition with tocilizumab in patients with rheumatoid arthritis (OPTION study): a double-blind, placebo-controlled, randomised trial

Chemokines in atherosclerosis: an update

The novel atherosclerosis locus at 10q11 regulates plasma CXCL12 levels

Protective role of CXC receptor 4/CXC ligand 12 unveils the importance of neutrophils in atherosclerosis

CXCR4 blockade induces atherosclerosis by affecting neutrophil function

A genome-wide association scan on the levels of markers of inflammation in sardinians reveals associations that underpin its complex regulation

A genome-wide association study for coronary artery disease identifies a novel susceptibility locus in the major histocompatibility complex

Physiology and pathophysiology of vascular signaling controlled by guanosine 3',5'-cyclic monophosphate-dependent protein kinase

Genome-wide association study in Han Chinese identifies four new susceptibility loci for coronary artery disease

Dysfunctional nitric oxide signalling increases risk of myocardial infarction

Rabbit models for the study of human atherosclerosis: from pathophysiological mechanisms to translational medicine

Generation of PPARgamma mono-allelic knockout pigs via zinc-finger nucleases and nuclear transfer cloning

Transgenic rabbits expressing human apolipoprotein a-I in the liver

Transgenic rabbits expressing human apolipoprotein(a) develop more extensive atherosclerotic lesions in response to a cholesterolrich diet

Human C-reactive protein does not promote atherosclerosis in transgenic rabbits

Genomewide association analysis of coronary artery disease

A genome-wide association study identifies PLCL2 and AP3D1-DOT1L-SF3A2 as new susceptibility loci for myocardial infarction in Japanese

Genomewide association identifies a susceptibility locus for coronary artery disease in the Chinese Han population

Genome-wide association of early-onset myocardial infarction with single nucleotide polymorphisms and copy number variants

SNPs on chromosome 5p15.3 associated with myocardial infarction in Japanese population

Genetic variants associated with Lp(a) lipoprotein level and coronary disease

Mechanisms of statin-induced myalgia assessed by physiogenomic associations

Sequence variants affecting eosinophil numbers associate with asthma and myocardial infarction

Genome-wide association study of coronary artery disease in the Japanese

Large-scale gene-centric analysis identifies novel variants for coronary artery disease

Acknowledgments This work was supported, in whole or in part, by National Institutes of Health Grants HL068878, HL117491, and HL129778 (to Y.E.C.), P50-HL117929 (to M.G.B.), and by the American Heart Association National Scientist Development Grant (15SDG24470155 to Y. G.).