Are hypertriglyceridemia and low HDL causal factors in the development of insulin resistance?

Atherosclerosis 233 (2014) 130e138

Contents lists available at ScienceDirect

Atherosclerosis journal homepage: www.elsevier.com/locate/atherosclerosis

Review

Are hypertriglyceridemia and low HDL causal factors in the development of insulin resistance? Naishi Li a, b, Jingyuan Fu c, Debby P. Koonen a, Jan Albert Kuivenhoven a, Harold Snieder d, Marten H. Hofker a, * a

University of Groningen, University Medical Center Groningen, Department of Molecular Genetics, Groningen, The Netherlands Department of Endocrinology, Key Laboratory of Endocrinology of Ministry of Health, Peking Union Medical College Hospital, Peking Union Medical College, Chinese Academy of Medical Science, Beijing, China c University of Groningen, University Medical Center Groningen, Department of Genetics, Groningen, The Netherlands d University of Groningen, University Medical Center Groningen, Department of Epidemiology, Genetic Epidemiology & Bioinformatics Unit, Groningen, The Netherlands b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 3 June 2013 Received in revised form 18 November 2013 Accepted 5 December 2013 Available online 7 January 2014

Insulin resistance often occurs with dyslipidemia as part of the metabolic syndrome and the current dominant paradigm is that insulin resistance leads to dyslipidemia. However, dyslipidemia may also cause insulin resistance; this was postulated 30 years ago, but has never been substantiated. Establishing whether dyslipidemia plays a causal role in the etiology of insulin resistance is important since it could reveal new avenues for combating type 2 diabetes. In this review we summarize recent evidence from epidemiological, genetic and intervention studies to re-address this old hypothesis. Ó 2014 Elsevier Ireland Ltd. All rights reserved.

Keywords: HDL-cholesterol Triglycerides Insulin resistance Dyslipidemia

Contents 1.

2. 3.

4.

5. 6.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .131 1.1. Need for new pharmaceutical targets to prevent type 2 diabetes (T2D) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131 1.2. Insulin-resistant state as target for preventing T2D . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131 1.3. Dyslipidemia as a possible causal factor of insulin resistance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131 Epidemiological studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .132 Genetic studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .133 3.1. Candidate gene studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133 3.2. Genome-wide association studies (GWAS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134 3.3. Mendelian randomization studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135 Intervention studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .135 4.1. LDL-C lowering treatments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135 4.2. Triglyceride lowering treatments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135 4.3. HDL modulating treatments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136 Using systems genetics to understand disease etiology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .136 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .136 Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137

* Corresponding author. Department of Molecular Genetics, University Medical Center Groningen, Hanzeplein 1, 9713 GZ Groningen, The Netherlands. Tel.: þ31 (0)50 363 5777; fax: þ31 (0)50 363 8971. E-mail addresses: [email protected], [email protected] (M.H. Hofker). 0021-9150/$ e see front matter Ó 2014 Elsevier Ireland Ltd. All rights reserved. http://dx.doi.org/10.1016/j.atherosclerosis.2013.12.013

N. Li et al. / Atherosclerosis 233 (2014) 130e138

Abbreviations ABCA1 ATP-binding cassette, sub-family A, member 1 BIP Bezafibrate Infarction Prevention trial CAD coronary artery disease CETP cholesterol ester transfer protein ER endoplasmic reticulum FAS fatty acid synthase FPG fasting plasma glucose GWAS genome-wide association studies HDL-C high-density lipoprotein cholesterol HOMA-IR homeostatic model assessment-insulin resistance IFG impaired fasting glucose IGT impaired glucose tolerance

1. Introduction 1.1. Need for new pharmaceutical targets to prevent type 2 diabetes (T2D) The prevalence of type 2 diabetes (T2D) is increasing so rapidly that it has become a serious problem worldwide [1,2]. Since we still have no ideal therapy for T2D or its chronic and serious complications, prevention is of the utmost importance. This point is clearly illustrated by several large-scale interventional trials in Chinese [3,4], American [5,6] and Finnish populations [7,8], which showed that a considerable proportion of patients with impaired glucose-tolerance (IGT) still developed T2D after several years of follow-up, in spite of apparently effective interventions such as lifestyle changes and metformin use. For instance, in the Da Qing Diabetes Prevention Outcome Study, 80% of individuals with IGT in the lifestyle intervention group (diet, or exercise, or diet plus exercise) had developed T2D after 20 years of follow-up, although this group did much better than the controls without intervention, in which 93% of IGT individuals suffered from T2D after 20 years [4]. These findings underscore the urgency for preventing T2D at a much earlier stage.

131

iNOS inducible nitric oxide synthase IR insulin resistance LDL-C low-density lipoprotein cholesterol LPL lipoprotein lipase MR Mendelian randomization NAFLD nonalcoholic fatty liver disease PPAR peroxisome proliferator-activated receptor rHDL reconstituted high density lipoproteins T2D type 2 diabetes TC total cholesterol TG triglycerides ULSAM Uppsala Longitudinal Study of Adult Men VLDL very low-density lipoprotein WOSCOPS West of Scotland Coronary Prevention Study

insulin resistance. In the 1960s, hypertriglyceridemia was found to be related to IR [10,11], but no compelling evidence for a causal relationship could be discovered. In 1982, Steiner and Vranic postulated the “hypertriglyceridemia-IR-hyperinsulinemia vicious cycle” [12], which emphasized that hypertriglyceridemia could cause IR, while IR and compensatory hyperinsulinemia would aggravate hypertriglyceridemia. They pointed out that it should be possible to prevent T2D by intervening to control hypertriglyceridemia. This was a pioneering hypothesis at that time, although it only concentrated on hypertriglyceridemia due to the then limited studies on blood lipid profiles. However, at the Banting lecture (American Diabetes Association, New Orleans) in 1988, Reaven coined the term “syndrome X” (later changed to the more popular “metabolic syndrome”) to describe the cluster of metabolic disorders including dyslipidemia, hypertension, hyperglycemia, and central obesity, with IR postulated as the common cause [13]. Since then, a lot of evidence has supported the dominant hypothesis that IR can induce dyslipidemia [14].

1.2. Insulin-resistant state as target for preventing T2D It is widely accepted that central adiposity leads to dyslipidemia, dysglycemia and hypertension, while hyperglycemia often appears with gradually impaired b-cell function (Fig. 1). Only many years later are the metabolic abnormalities followed by T2D, characterized by marked loss of b-cell function. This typical sequence of pathophysiological processes happens during the natural disease course in most patients with T2D. However, T2D is a heterogeneous disease, hence the stage of normal glucose tolerance but with insulin resistance (IR) is the potential target period for prevention, before IGT develops in such individuals. But what can we do for individuals with normal glucose tolerance but an insulin-resistant state? To provide better prevention, we must first better understand the etiology of insulin resistance. 1.3. Dyslipidemia as a possible causal factor of insulin resistance The concept of insulin resistance was first postulated in the 1930s [9]. Two types of diabetic patients were differentiated by their blood sugar response to insulin therapy, and the description of the insulin-insensitive type represents the earliest research into

Fig. 1. Future target for intervention to prevent type 2 diabetes. Current intervention takes place in IFG/IGT (prediabetes) stage, around 5 years before type 2 diabetes develops. We believe that the stage of mild insulin resistance but normal glucose tolerance, earlier than IFG/IGT stage, will be a better phase of invention for preventing type 2 diabetes (bottom panel). This phase is further characterized by increased plasma triglycerides and low HDL-C (top panel).

132

N. Li et al. / Atherosclerosis 233 (2014) 130e138

Compensatory hyperinsulinemia due to IR can induce increased flux of free fatty acids, raise production of triglycerides in the liver, and decrease high-density lipoprotein cholesterol (HDL-C). Thus, by the 1990s, dyslipidemia was believed to be a bystander rather than a causal factor. Several years passed and knowledge of dyslipidemia increased dramatically. Recently, new evidence, including results from genome-wide association studies (GWAS), has revealed support for the old “vicious cycle” hypothesis [15,16]. However, we suggest replacing “hypertriglyceridemia” with “dyslipidemia” as it is not yet known which types of dyslipidemia lead to insulin resistance. We propose a revision of the old hypothesis to state: “dyslipidemia is a causal factor of insulin resistance”. Here we should make it clear that we are focusing on “dyslipidemia” rather than “lipotoxicity”. The latter term, which was coined by Unger nearly 20 years ago [17,18], means ectopic lipid deposition in the cells of non-adipose tissues such as liver, muscle, and heart [19]. In recent years, lipotoxicity has been accepted as one of the main mechanisms behind IR. Although lipotoxicity is associated with hypertriglyceridemia to some extent, they are not one and the same disorder (Box 1). We focus on “dyslipidemia” because we may be able to prevent T2D in the early insulin-resistant stage by alleviating dyslipidemia (Fig. 1). Could all types of dyslipidemia be possible causal factors of IR? This is unlikely, and only triglyceride and HDL-C are potential candidates, rather than LDL-C. This is based on the Framingham Heart Study, which showed that hypertriglyceridemia and low HDL-C were more prevalent in T2D patients than in the normal population, while high LDL-C did not differ significantly in these two groups [20]. This suggested that hypertriglyceridemia and low HDL-C, but not high LDL-C, might be causal factors of T2D. Various types of studies, including some GWAS-related research (shown below) support this inference (Fig. 2). In addition to genetic and epidemiological studies, plausible biological mechanisms have also emerged supporting a role for HDL in insulin resistance and diabetes. These have been recently reviewed [15,16,21]. Importantly HDL can improve insulin secretion from bcells, and increase insulin sensitivity in the target tissues such as adipose tissue, muscle and liver. In addition, HDL has a positive effect on b-cell survival and insulin secretion. Some of the benign effects of HDL are due to its capability to suppress inducible nitric oxide synthase (iNOS) and fatty acid synthase (FAS). Other effects can be ascribed to the role of HDL on cholesterol homeostasis. In experimental models, it has been shown that HDL, ABCA1 and ABCG1 play a role in the cholesterol homeostasis of b-cells. The principal apolipoprotein moiety of HDL, APOA-I, was found to induce the phosphorylation of AMP kinase (AMPK), thereby facilitating glucose uptake into

Box 1 Differences between dyslipidemia and lipotoxicity.

Definition Location Quantitative evaluation Clinical marker Causing insulin resistance Clinical use

Dyslipidemia

Lipotoxicity

Abnormal serum lipoprotein concentrations Circulation Blood test

Ectopic lipid accumulation Liver and muscle tissue MRI

Yes Possibly

No Yes

Easy

Difficult

myocytes and inhibiting gluconeogenesis in hepatocytes [15]. Additional evidence is emerging that HDL regulates fat storage in adipocytes. These properties of HDL indicate that it is capable of playing a role in both insulin resistance as well as in b-cell function. There is considerable heterogeneity among the HDL subclasses, ranging from small to large HDL. In general, the larger HDL classes show the more beneficial effects on insulin sensitivity [22,23]. The strong inverse relationship between HDL and VLDL (very lowdensity lipoprotein) levels implies that VLDL plays a role as well. VLDL will have indirect effects, but may also influence insulin sensitivity directly. The mechanism through which hypertriglyceridemia and low HDL-C could lead to IR still needs to be resolved. Lipotoxicity, inflammation and endoplasmic reticulum (ER) stress are three widely-accepted mechanisms known to induce IR [24e27]. Some human studies have suggested that high serum triglycerides seem to provoke a considerable amount of fatty acids to be deposited in cells in ectopic lipid storage [28]. Lipotoxicity could thus be the main causal mechanism following dyslipidemia [16,27,29]. However, it is also possible that dyslipidemia is a direct cause of inflammation, ER stress, or other mechanisms leading to IR in the absence of lipotoxicity. However, mechanisms whereby IR causes increased VLDL and plasma free fatty acids (FFA) levels are abundantly documented. IR stimulates lipogenesis and cholesterol synthesis. Hence, excess amounts of FFA are generated in adipose tissue and transferred to the liver, which results in the overproduction of VLDL, increased plasma triglycerides and a reduction of HDL-C levels through the action of CETP (cholesterol ester transfer protein) and increased clearance of HDL-C by kidney (Fig. 2) [14,30]. The role of LDL in IR is less clear and if it does play a role, it will be a very modest one. However, while LDL does not seem to modulate insulin sensitivity, it is capable of affecting b-cell function, as it has been shown that high levels of plasma LDL promote loss of b-cells [31,32]. Thus, dyslipidemia and IR may have a reciprocal, positive feedback relationship, and they may reinforce each other. Hypertriglyceridemia and low HDL-C are associated with IR, while the consequences of IR involve dyslipidemia. This “vicious cycle” accelerates the development of IR in the early stages of the development of T2D. In this review, we focus on the evidence from human studies that supports the hypothesis that hypertriglyceridemia and low HDL-C may be causal factors of insulin resistance. 2. Epidemiological studies If hypertriglyceridemia and low HDL-C are causal factors in the development of IR, one would expect the two types of dyslipidemia to be able to predict future insulin sensitivity. Several long-term prospective studies have been carried out in line with this idea. Although these studies cannot be seen as rigorous evidence, they do support a potentially causal role of HDL. One study is the Uppsala Longitudinal Study of Adult Men (ULSAM) showing that low HDL-C is a long-term predictor of insulin sensitivity [33]. The authors investigated 770 men who had participated in a health survey 20 years earlier and determined their insulin sensitivity through hyperinsulinemic euglycemic clamps. After adjusting for BMI as a confounder, low HDL-C levels proved an independent predictor of IR 20 years later, while triglycerides were not. There have been other longitudinal studies showing that hypertriglyceridemia, low HDL-C levels and/or the serum triglyceride/HDL-C index can predict the future risk of developing T2D [34e36]. In the Framingham offspring study [37], 1004 offspring with baseline pre-diabetes were followed up for seven years and 118 new-onset T2D cases were recorded. Low HDL-C was proved to be one of the key non-glycemic traits that predicted later T2D in pre-diabetes (IGT or impaired fasting glucose, IFG) individuals. Although it is unfortunate that the latter study did not measure

N. Li et al. / Atherosclerosis 233 (2014) 130e138

133

Fig. 2. Schematic representation of the relationship between insulin resistance, hypertriglyceridemia and low HDL cholesterol. 2A) The mechanisms by which insulin resistance leads to hypertriglyceridemia and low HDL cholesterol. Insulin resistance induces compensatory hyperinsulinemia, causing overproduction and secretion of VLDL in the liver, which leads to increased plasma VLDL-TG. At the same time, insulin resistance activates the process of lipolysis, which is originally inhibited by insulin, so that FFAs are produced and released to blood stream. FFAs also can increase production of triglyceride as a kind of substrate in the liver by entering hepatocytes, which aggravates hypertriglyceridemia. After bidirectional transfer of CETP, the HDL particles without CE are susceptible to be breakdown and cleared by kidney, causing decreased HDL levels. 2B) The mechanisms leading from dyslipidemia to insulin resistance is mostly supported by genetic variants in lipid genes, depicted on top, leading to increased serum triglycerides or decreased HDL-C. It is therefore likely that lipoproteins can induce insulin resistance. The mechanism for this relationship is not known, but is likely to involve several steps. In addition, we cannot rule out that the genetic variants cause pleiotropic effects leading to insulin resistance, unrelated to the lipoproteins.

insulin sensitivity as an outcome, we speculate that the deteriorating IR could be a primary cause of the ensuing decline in b cell function. Although it is not possible to derive causal insights from these epidemiological studies, the combined results are in line with the proposition that hypertriglyceridemia and/or low HDL-C play a role in the onset of IR, and can thus be used as long-term predictors of insulin sensitivity.

3. Genetic studies 3.1. Candidate gene studies Three representative genes which play a role in VLDL metabolism encode proteins in lipoprotein pathways that are also known to be involved in IR, thereby supporting a causal link

134

N. Li et al. / Atherosclerosis 233 (2014) 130e138

between dyslipidemia and IR. These three genes are the lipoprotein lipase (LPL), apolipoprotein C3 (ApoC3) and angiopoietin-like 3 (ANGPTL3), which are involved in the lipolysis of triglycerides, thereby mediating the uptake of fatty acids by the target tissues. LPL is bound to the proteoglycans on the endothelial lining of the vessel wall, where large triglyceride rich lipoproteins also anchor and will become subject to lipolysis of their triglyceride moiety. LPL is crucial in the conversion of chylomicrons and VLDL into their remnant particles. ApoC3 is bound to the lipoproteins and inhibits the action of LPL. Thus, ApoC3 can control the lipolysis rate of lipoproteins. Loss-of-function alleles in mice and humans lower plasma triglyceride levels [38]. ANGPTL3 also encodes an inhibitor of LPL [39]. Several studies involving mutations in or around LPL, ApoC3 and ANGTL3 have provided insight into the link of triglyceride metabolism and insulin signaling. In a case report of a familial lipoprotein lipase (LPL) deficiency, two affected sisters carried loss-of-function mutations in the LPL gene and they both suffered from insulin-resistant diabetes and severe hypertriglyceridemia. After their dyslipidemia was successfully treated by bilio-pancreatic diversion, the patients recovered from their diabetes and oral glucose-tolerant tests showed completely normal results [40]. In addition, the glucose infusion rate of hyperinsulinemiceeuglycemic clamp studies increased dramatically after surgery, indicating elevated insulin sensitivity. In another study, 85 heterozygous genetic carriers of either Gly188Glu mutation or a splice site mutation (C–>A in position 3 at the acceptor splice site of intron 6) in the LPL gene, which both result in a catalytically inactive product, were compared with 108 unaffected subjects from the same families. The results showed that HOMA-IR (homeostatic model assessment-insulin resistance) index values were significantly higher in carriers than in non-carriers [41], indicating an attenuated sensitivity to insulin in the mutant carriers. Furthermore, based on 12 single nucleotide polymorphisms in the 30 end region of the LPL gene, which may change serum triglyceride levels by regulating LPL activity, the risk haplotypes for insulin resistance were discovered and replicated [42]. In hyperinsulinemiceeuglycemic clamp studies in 291 individuals from a large, family-based population, the same risk haplotypes were found to be related to the glucose infusion rate [43]. These studies thus provide evidence that loss of LPL gene function leads to impaired insulin sensitivity. This is likely mediated through its effect on plasma triglyceride levels as well as lipid flux into the tissues. Another important gene encoding a protein in the pathway controlling VLDL-TG levels is ApoC3, which is related to IR. In a study of 95 Asian Indian men, the T-455C and C-482T polymorphisms of the ApoC3 gene were found to be strongly associated with hypertriglyceridemia [44]. Moreover, it was found that the prevalence of non-alcoholic fatty liver disease (NAFLD) was 38% among variant-allele carriers, but absent among wild-type homozygotes (P < 0.001). Furthermore, those individuals with NAFLD were more insulin-resistant than the individuals without it (insulin-sensitivity index, 2.0  2.0 vs. 3.5  1.6; P < 0.001), consistent with the strong correlation between NAFLD and IR. The authors concluded that the polymorphisms C-482T and T-455C in ApoC3 are associated with NAFLD and IR. Because ApoC3 is known as a protein-inhibiting lipolysis of VLDL, the main effect of these ApoC3 variants (which lead to raised circulating ApoC3 levels) is a rise in VLDL-TG. Thus the present study greatly supports our proposed role for VLDL-TG in the etiology of IR. Complete ANGPTL3 deficiency is associated with increased activity of LPL accelerating hydrolysis of VLDL and leading to improved insulin sensitivity [45]. As the cause of type 2 familial hypobetalipoproteinemia, research into ANGPTL3 also supports our hypothesis that hypertriglycemia and low HDL-C can be causal factors of insulin resistance.

Two other genes affecting insulin signaling play a role in HDL metabolism. These two genes encode cholesterol ester transfer protein (CETP) and hepatic lipase (LIPC). CETP is a glycoprotein which mediates the transfer of cholesterol ester from HDL to apolipoprotein B-containing lipoproteins, and of triglyceride from VLDL to LDL and HDL. LIPC is an enzyme which can catalyze the hydrolysis of triglyceride from TG-rich lipoproteins. Hence, both are important in dyslipidemia. Taq 1B, which is located on the first intron of CETP, is associated with the level of activity of CETP. In a population from the Canary Islands, polymorphisms were found to be related to both circulating lipid levels and IR. Individuals with the B1B1 genotype had significantly lower circulating HDL cholesterol and higher HOMA-IR levels [46]. Similarly, there is evidence that the functional C-514 T or G-250 A polymorphisms in the promoter of LIPC also affect both circulating lipid levels and IR [47e49]. Although it is unclear how these two gene polymorphisms induce insulin resistance, it is likely that the effect is mediated through lipoproteins. In conclusion, genetic data on LPL, ApoC3, ANGPTL3, CETP and LIPC convincingly show that while the primary cause of the mutations results in dyslipidemia, the secondary consequences lead to IR. In the absence of a direct role for LPL, Apoc3 or ANGPTL3 in insulin resistance, these data support a causal role for dyslipidemia in the etiology of IR. Two examples of genes controlling lipid metabolism and lipid handling are the ATP-binding cassette, sub-family A (ABCA1) and Lamin A/C (LMNA), respectively. These genes also affect glucose metabolism, but through other mechanisms. ABCA1 has been discovered in HDL-deficient Tangier patients [50]. It is crucial in the formation of HDL and also plays a role in reverse cholesterol transport. In a Dutch cohort, it was discovered that heterozygous carriers of a loss-of-function mutation show lower glucose tolerance and decreased insulin secretion [51]. However, the patients do not show altered insulin sensitivity. The defect is mainly confined to a loss of b-cell function. Thus, the effect of the ABCA1 mutation on glucose homeostasis is unrelated to differences in plasma HDL. A more general pattern of disturbed lipoprotein metabolism is seen in LMNA-deficient subjects showing lipodystrophy [52]. These patients show a loss of adipose tissue development. In the absence of a depot for fat storage, the patients develop a marked increase in plasma triglyceride and a decrease in HDL, as well as severe insulin resistance, which precedes the development of T2D. Although these patients are ideal for testing the potential role of lipoproteins in the development of IR, the sequence of the events is still under debate [52e54]. 3.2. Genome-wide association studies (GWAS) Knowledge on the genetics of complex diseases has developed rapidly in recent years, especially since the completion of the Human Genome Project. The development of highly dense genotyping arrays has led to great success in identifying genetic variants using GWAS. So far, about 100 loci related to dyslipidemia [55,56] have been reported and at least nine loci for IR have been proposed [57]. Interestingly, several susceptibility loci have been found to be shared by both traits. For instance, both rs972283 for IR and rs4731702 for HDL-C are located near the KLF14 gene in 7q32.3 and are closely linked. In addition, rs11642841, which is located in the region of the well-known FTO gene for obesity and T2D, is closely linked with rs1421085 for the HDL-C trait. These might suggest common pathways for mechanisms involved in IR and hypertriglyceridemia/low HDL-C, which would be consistent with our hypothesis.

N. Li et al. / Atherosclerosis 233 (2014) 130e138

3.3. Mendelian randomization studies Genetic association studies will not resolve what mechanism links SNP information to dyslipidemia and IR. For this purpose, the Mendelian randomization (MR) approach has been developed [58,59]. MR is suitable to investigate the causal relationship between clinical observations (biomarker or intermediate/endophenotypes) and disease outcomes using genetic information. The basic principle of MR is that if a biomarker (i.e., dyslipidemia in our hypothesis) is a causal factor for the disease outcome (i.e., insulin resistance in our hypothesis), the genetic variation underlying this biomarker should also be associated with the disease. The association should be consistent with established epidemiological relationships, i.e., SNP alleles that increase triglycerides or decrease HDL-C should be associated with an increased risk of IR. As genotypes are assigned randomly at birth when transmitted from parents to offspring during meiosis, MR is believed to act as a “natural randomized clinical trial”. As such, MR studies can help to test whether the association between a risk marker and outcome is causal, or whether the marker is raised as part of the disease process (reverse causation) [60]. Following this approach, ten common genetic variants robustly associated with circulating triglyceride levels were studied in 5637 T2D patients and 6860 controls [61]. There was no evidence that carriers of greater numbers of triglyceride-raising alleles had an increased risk of T2D or higher levels of fasting-insulin or fasting-glucose. The data thus implied that hypertriglyceridemia was not a causal factor of IR [61]. In contrast, another MR study drew the opposite conclusion e that genetic predisposition to low HDL-C or high triglycerides was related to elevated T2D risk [62]. However, both studies had some limitations. The first study covered a limited number of genetic variants (10 loci accounting for only 3e5% of the variation in circulating triglycerides), while the second study covered 25 loci, jointly explaining nearly 10% of the genetic variation, but insulin resistance was not tested. A carefully designed MR study is needed to clarify the causal relationship between dyslipidemia and IR, and should include all the established genetic loci of dyslipidemia, accurate phenotyping of IR cases, and a large sample size. 4. Intervention studies Over the last few decades, many lipid-modulating drugs have been introduced into clinical medicine. Classified according to their main target, these reduce levels of LDL-C or triglycerides, while HDL-C raising drugs have been developed recently. In line with our hypothesis, we would expect the latter two to alleviate the development of insulin resistance, while the first would not. Unfortunately, since most clinical trials have studied the effectiveness and safety of the drugs in relation to cardiovascular outcomes, data on patients’ insulin sensitivity are limited to a few well-documented trials, which we will discuss below. 4.1. LDL-C lowering treatments So far, statins have been proven to lower coronary artery disease (CAD) risk effectively by reducing serum LDL-C. Their effects on serum triglycerides or HDL-C are moderate. However, over the last ten years, statin use has been reported to lead to an increase in the incidence of T2D. In 2010, a large meta-analysis included 13 statin trials with 91,140 participants [63] and confirmed that statin therapy was indeed associated with a slightly increased risk of developing T2D. Another meta-analysis included 16 trials with 1146 participants and focused on insulin sensitivity, comparing the effects of different statins [64]. They found that pravastatin significantly improved insulin sensitivity, whereas simvastatin made it

135

significantly worse. These results were consistent with earlier data from the West of Scotland Coronary Prevention Study (WOSCOPS) [65], where the authors concluded that the assignment to pravastatin therapy resulted in a 30% reduction of developing diabetes. Without having any data on insulin sensitivity, these authors speculated that pravastatin therapy might, at least partly, reduce the development of diabetes by lowering plasma triglyceride levels by as much as 12%. Pitavastatin is a potent novel statin with LDL-C lowering effects comparable to those of atorvastatin, but with some HDL-C raising effects (5.9e8.2% in a general cohort and up to 22.4% in patients with low circulating HDL-C) [66]. In the LIVES study (n ¼ 308), pitavastatin was found to significantly decrease HbA1c in diabetic patients [67]. The authors speculated that pitavastatin might raise HDL-C and thereby improve insulin sensitivity. Ezetimibe, an intestinal cholesterol absorption inhibitor, can decrease LDL-C by approximately 20%, triglycerides by up to 8%, and raises HDL-C by 1e4%. A small clinical study with 75 participants showed that ezetimibe monotherapy for 12 weeks effectively decreased HOMA-IR, along with improvement of serum LDL-C, triglyceride, and HDL-C levels [68]. Another small study with 45 patients investigated the long-term effect of ezetimibe in patients of NAFLD and found HOMA-IR and other abnormalities of NAFLD also improved significantly [69]. However, much larger clinical studies are needed to validate these effects and more in-depth studies should investigate whether insulin sensitivity is improved through controlling dyslipidemia by LDL-C lowering therapies. More indirect effects on lipoprotein metabolism are being exerted by bile acid sequestrants. One such sequestrant, colesevelam has been recently reviewed [70]. Strikingly, colesevelam, which mainly acts on the removal of intestinal bile acids in the gut, leads to lower plasma cholesterol and LDL-C levels [71]. Recently, it has been recognized that colesevelam also shows marked beneficial effects on glycemic control, when used in combination with other antidiabetic drugs [72]. Unfortunately, further studies showed that colesevelam could not alleviate IR [73e75]. The mechanism of colesevelam might be very complex, while its effect of increasing triglyceride levels might produce some adverse effects on insulin sensitivity [74]. Thus, some LDL-C lowering drugs act favorably by reducing HbA1c levels or delaying the onset of diabetes. However, this beneficial effect appears to be inconsistent amongst the LDL-C lowering drugs. Understanding the mechanisms behind the effects of statins on insulin sensitivity is difficult, because statins not only decrease plasma lipids but also have other effects. First, statins can reduce chronic low-grade inflammation, which could lead to improved insulin sensitivity [65]. Second, statins may adversely affect b-cell function, i.e., while LDL-C lowering may be beneficial for insulin resistance, they could negatively impact b-cell function. Third, differences between the various statins in terms of efficacy and biodistribution could lead to different outcomes. This could explain differences in response to the different drugs and will be of great interest to better understanding the mechanisms of IR. 4.2. Triglyceride lowering treatments Fibric acid derivatives (fibrates), which are peroxisome proliferator-activated receptor (PPAR) alpha agonists, reduce serum triglyceride by 25e50% and raise HDL-C by 10e20%, while reducing LDL-C by 5e20%. Bezafibrate was found to be beneficial for delaying the development of pre-diabetes into diabetes in 1990 [76]. In subgroups of the Bezafibrate Infarction Prevention (BIP) trial with impaired fasting glucose (n ¼ 303), incidence of diabetes was seen to be lower in the bezafibrate treatment group than in controls [77]. In another subgroup of the BIP trial, a group of patients suffering

136

N. Li et al. / Atherosclerosis 233 (2014) 130e138

from coronary artery disease treated by bezafibrate (n ¼ 1262) had significantly lower HOMA-IR values than the control group (n ¼ 1242) after two years of follow-up [78]. Thus, treatment with bezafibrate delayed onset of IR over two years. In addition, it was shown in a group of 351 patients that treatment with bezafibrate can improve both IR (HOMA-IR) and b-cell function [79]. However, other studies on fibrates (including gemfibrozil, fenofibrate and bezafibrate) reached paradoxical conclusions [80,81]. Some research suggested that the mechanism via which PPAR alpha agonists improve IR was very complex and further basic studies are needed. Interpretation is complicated, as fibrates are shown to increase b-oxidation and thus have effects that improve insulin sensitive that are different from their effects on plasma lipids. It is therefore still too early to say whether triglyceride lowering treatment could delay insulin resistance, in spite of some supporting evidence. 4.3. HDL modulating treatments Treatments modulating HDL metabolism are currently under development but only data from small-scale studies are available. In a recent double-blind, crossover study, reconstituted high density lipoprotein (rHDL) infusion in 13 patients was shown to lower plasma glucose levels significantly, while plasma insulin levels increased after 2.5 hours’ infusion [82]. Meanwhile, the HOMA-IR measurements in the rHDL and control groups showed no difference. The reduction of plasma glucose could not be explained only by enhanced b-cell function, because the reduction appeared before the elevation of plasma insulin levels. In skeletal muscle cells it was demonstrated that the AMPK pathway was activated by rHDL [82], which indirectly improved insulin sensitivity of the muscles. Interestingly, CETP inhibitors showed significant beneficial effects on IR in the ILLUMINATE trial (n ¼ 6661), and were probably related to elevated serum HDL-C levels [83]. In summary, intervention studies have shown that insulin sensitivity is improved by therapies aiming to lower triglycerides and to modulate HDL, but robust insight into the exact mechanisms is still lacking. 5. Using systems genetics to understand disease etiology One of the major promises at the moment is that systems genetics approaches will lead to much needed insight into the etiology of IR. In particular, because GWAS studies still leave a large proportion of the heritability unexplained and there is hot debate on the sources of this “missing heritability”. These limitations severely hamper the generation of robust mechanistic insights based on causality. Moreover, the current set of GWAS-based genetic loci has only limited power to deliver functional insight. It should be realized that most of the associated SNPs are not in coding regions and they are more likely to be involved in regulatory processes. With the greatly extended regulatory knowledge gained from ENCODE (the Encyclopedia of DNA Elements, http://genome. ucsc.edu/ENCODE/), the lead associated SNPs are, in most cases, in linkage disequilibrium with unknown causal variants [84]. In addition, causal variants may affect the expression of a nearby gene or genes, but there is still a lot of uncertainty about the identity of the gene that is the subject of the causal variation. Also, the causal variant may regulate genes at greater distances or even on different chromosomes. These limitations reduce the power of a Mendelian randomization study, while the lack of understanding of the precise genetic changes further limit our understanding of the disease etiology. Systems genetics approaches seek to understand complex disease traits by integrating systems biology approaches with

those of genetics. Thus, in the post-GWAS era, these approaches are becoming a powerful way of inferring causal roles from genotype to phenotypes. To understand the causal relationship between dyslipidemia and IR, we have to develop novel approaches, such as identifying the causal variants and genes of dyslipidemia from ENCODE and other bioinformatic data, treating those genetic variants as multifactorial perturbations in biological systems, and examining their effect on all aspects and traits related to IR. Subsequently, we should perform functional studies in model organisms and longitudinal studies in biological systems to confirm the findings (Box 2). 6. Conclusion Unraveling whether dyslipidemia is a possible cause of insulin resistance is important, as it has many ramifications regarding the early diagnosis and prevention of type 2 diabetes. LDL-C has been excluded as a relevant factor, since there is no evidence to support a causal relationship with insulin resistance. However, the role of statins should be explored further as there are many indications that these drugs have effects beyond reducing LDL that impact on the susceptibility to develop diabetes. Hypertriglyceridemia and low HDL-C have been shown to be independent, long-term predictors of insulin sensitivity in longitudinal studies, while some studies on genes related to hypertriglyceridemia and/or low HDL-C have revealed that these genes are also linked to insulin sensitivity. Furthermore, controlling hypertriglyceridemia and low HDL-C has delayed the onset of insulin resistance in some intervention studies. Additional genetic, Mendelian randomization and systems biology studies are now feasible and needed to confirm the causal role of hyperlipidemia in insulin resistance and its underlying mechanisms. Such studies will lead to urgently needed insight into ways to prevent insulin resistance and the subsequent development of type 2 diabetes. Acknowledgments We thank J.L. Senior for critically reading the manuscript. There are no conflicts of interest. This work was performed within the framework of CTMM, the Center for Translational Molecular Medicine (www.ctmm.nl), project PREDICCt (grant 01C-104), and supported an NWO VENI grant 863.09.007 (JF), by the Netherlands Heart Foundation, Dutch Diabetes Research Foundation and Dutch Kidney Foundation. The Graduate School for Drug Exploration (GUIDE), University of Groningen also financially supported this work. Box 2 Looking to the future: what research should be done to test our hypothesis?

1. We need longitudinal studies on large groups of dyslipidemia patients, especially on patients with monogenetic dyslipidemia, and intervention studies on dyslipidemia with insulin sensitivity as the main outcome. 2. We need carefully designed Mendelian randomization studies to clarify the causal relationship between dyslipidemia and insulin resistance, preferably including research in diverse ethnic groups. 3. Further elucidation of the underlying etiology of insulin resistance will require a comprehensive approach involving study of genetic variation, intervention of environmental factors, systematic profiling of “omics” data, advanced statistical models, and functional studies in model organisms.

N. Li et al. / Atherosclerosis 233 (2014) 130e138

References [1] Meigs JB. Epidemiology of type 2 diabetes and cardiovascular disease: translation from population to prevention: the Kelly West award lecture 2009. Diabetes Care 2010;33:1865e71. [2] Chen L, Magliano DJ, Zimmet PZ. The worldwide epidemiology of type 2 diabetes mellitus e present and future perspectives. Nat Rev Endocrinol 2012;8:228e36. [3] Pan XR, Hu YH, Li GW, Liu PA, Bennett PH, Howard BV. Impaired glucose tolerance and its relationship to ECG-indicated coronary heart disease and risk factors among Chinese. Da Qing IGT and Diabetes Study. Diabetes Care 1993;16:150e6. [4] Li G, Zhang P, Wang J, et al. The long-term effect of lifestyle interventions to prevent diabetes in the China Da Qing Diabetes Prevention Study: a 20-year follow-up study. Lancet 2008;371:1783e9. [5] Knowler WC, Barrett-Connor E, Fowler SE, et al. Reduction in the incidence of type 2 diabetes with lifestyle intervention or metformin. N Engl J Med 2002;346:393e403. [6] Knowler WC, Fowler SE, Hamman RF, et al. 10-year follow-up of diabetes incidence and weight loss in the Diabetes Prevention Program Outcomes Study. Lancet 2009;374:1677e86. [7] Lindstrom J, Ilanne-Parikka P, Peltonen M, et al. Sustained reduction in the incidence of type 2 diabetes by lifestyle intervention: follow-up of the Finnish Diabetes Prevention Study. Lancet 2006;368:1673e9. [8] Tuomilehto J, Lindstrom J, Eriksson JG, et al. Prevention of type 2 diabetes mellitus by changes in lifestyle among subjects with impaired glucose tolerance. N Engl J Med 2001;344:1343e50. [9] Himsworth H. Diabetes mellitus: its differentiation into insulin sensitive and insulin insensitive types. Lancet 1936;227:127e30. [10] Reaven GM, Lerner RL, Stern MP, Farquhar JW. Role of insulin in endogenous hypertriglyceridemia. J Clin Invest 1967;46:1756e67. [11] Ford Jr S, Bozian RC, Knowles Jr HC. Interactions of obesity, and glucose and insulin levels in hypertriglyceridemia. Am J Clin Nutr 1968;21:904e10. [12] Steiner G, Vranic M. Hyperinsulinemia and hypertriglyceridemia, a vicious cycle with atherogenic potential. Int J Obes 1982;6(Suppl. 1):117e24. [13] Reaven GM. Banting lecture 1988. Role of insulin resistance in human disease. Diabetes 1988;37:1595e607. [14] Reaven G. Insulin resistance and coronary heart disease in nondiabetic individuals. Arterioscler Thromb Vasc Biol 2012;32:1754e9. [15] von Eckardstein A, Sibler RA. Possible contributions of lipoproteins and cholesterol to the pathogenesis of diabetes mellitus type 2. Curr Opin Lipidol 2011;22:26e32. [16] Drew BG, Rye KA, Duffy SJ, Barter P, Kingwell BA. The emerging role of HDL in glucose metabolism. Nat Rev Endocrinol 2012;8:237e45. [17] Lee Y, Hirose H, Ohneda M, Johnson JH, McGarry JD, Unger RH. Beta-cell lipotoxicity in the pathogenesis of non-insulin-dependent diabetes mellitus of obese rats: impairment in adipocyte-beta-cell relationships. Proc Natl Acad Sci U S A 1994;91:10878e82. [18] Unger RH. Lipotoxicity in the pathogenesis of obesity-dependent NIDDM. Genetic and clinical implications. Diabetes 1995;44:863e70. [19] Unger RH, Scherer PE. Gluttony, sloth and the metabolic syndrome: a roadmap to lipotoxicity. Trends Endocrinol Metab 2010;21:345e52. [20] Wilson PW, Kannel WB, Anderson KM. Lipids, glucose intolerance and vascular disease: the Framingham Study. Monogr Atheroscler 1985;13:1e11. [21] Mineo C, Shaul PW. Novel biological functions of high-density lipoprotein cholesterol. Circ Res 2012;111:1079e90. [22] Filippatos TD, Rizos EC, Tsimihodimos V, Gazi IF, Tselepis AD, Elisaf MS. Small high-density lipoprotein (HDL) subclasses are increased with decreased activity of HDL-associated phospholipase A(2) in subjects with prediabetes. Lipids 2013;48:547e55. [23] von Eckardstein A, Nofer JR. Frail HDLs and stiff arteries in type 2 diabetes in juveniles. Diabetes 2013;62:2662e4. [24] Samuel VT, Shulman GI. Mechanisms for insulin resistance: common threads and missing links. Cell 2012;148:852e71. [25] Chavez JA, Summers SA. A ceramide-centric view of insulin resistance. Cell Metab 2012;15:585e94. [26] Glass CK, Olefsky JM. Inflammation and lipid signaling in the etiology of insulin resistance. Cell Metab 2012;15:635e45. [27] Kraegen EW, Cooney GJ, Turner N. Muscle insulin resistance: a case of fat overconsumption, not mitochondrial dysfunction. Proc Natl Acad Sci U S A 2008;105:7627e8. [28] Bickerton AS, Roberts R, Fielding BA, et al. Preferential uptake of dietary fatty acids in adipose tissue and muscle in the postprandial period. Diabetes 2007;56:168e76. [29] Ikonen E. Cellular cholesterol trafficking and compartmentalization. Nat Rev Mol Cell Biol 2008;9:125e38. [30] Cornier MA, Dabelea D, Hernandez TL, et al. The metabolic syndrome. Endocr Rev 2008;29:777e822. [31] Brunham LR, Kruit JK, Verchere CB, Hayden MR. Cholesterol in islet dysfunction and type 2 diabetes. J Clin Invest 2008;118:403e8. [32] Rutti S, Ehses JA, Sibler RA, Prazak R, Rohrer L, Georgopoulos S. Low- and highdensity lipoproteins modulate function, apoptosis, and proliferation of primary human and murine pancreatic beta-cells. Endocrinology 2009;150: 4521e30.

137

[33] Riserus U, Arnlov J, Berglund L. Long-term predictors of insulin resistance: role of lifestyle and metabolic factors in middle-aged men. Diabetes Care 2007;30: 2928e33. [34] Wilson PW, Meigs JB, Sullivan L, et al. Prediction of incident diabetes mellitus in middle-aged adults: the Framingham Offspring Study. Arch Intern Med 2007;167:1068e74. [35] Gupta AK, Dahlof B, Dobson J, Sever PS, Wedel H, Poulter NR. Determinants of new-onset diabetes among 19,257 hypertensive patients randomized in the Anglo-Scandinavian Cardiac Outcomes trial e blood pressure lowering arm and the relative influence of antihypertensive medication. Diabetes Care 2008;31:982e8. [36] Giannini C, Santoro N, Caprio S, Kim G, Lartaud D, Shaw M. The triglycerideto-HDL cholesterol ratio: association with insulin resistance in obese youths of different ethnic backgrounds. Diabetes Care 2011;34:1869e74. [37] Wilson PW, D’Agostino RB, Fox CS, Sullivan LM, Meigs JB. Type 2 diabetes risk in persons with dysglycemia: the Framingham Offspring Study. Diabetes Res Clin Pract 2011;92:124e7. [38] Hofker MH. APOC3 null mutation affects lipoprotein profile APOC3 deficiency: from mice to man. Eur J Hum Genet 2010;18:1e2. [39] Arca M, Minicocci I, Maranghi M. The angiopoietin-like protein 3: a hepatokine with expanding role in metabolism. Curr Opin Lipidol 2013;24:313e20. [40] Mingrone G, Henriksen FL, Greco AV, et al. Triglyceride-induced diabetes associated with familial lipoprotein lipase deficiency. Diabetes 1999;48: 1258e63. [41] Holzl B, Iglseder B, Sandhofer A, et al. Insulin sensitivity is impaired in heterozygous carriers of lipoprotein lipase deficiency. Diabetologia 2002;45: 378e84. [42] Goodarzi MO, Taylor KD, Guo X, et al. Haplotypes in the lipoprotein lipase gene influence fasting insulin and discovery of a new risk haplotype. J Clin Endocrinol Metab 2007;92:293e6. [43] Goodarzi MO, Guo X, Taylor KD, et al. Lipoprotein lipase is a gene for insulin resistance in Mexican Americans. Diabetes 2004;53:214e20. [44] Petersen KF, Dufour S, Hariri A, et al. Apolipoprotein C3 gene variants in nonalcoholic fatty liver disease. N Engl J Med 2010;362:1082e9. [45] Robciuc MR, Maranghi M, Lahikainen A, et al. Angptl3 deficiency is associated with increased insulin sensitivity, lipoprotein lipase activity, and decreased serum free fatty acids. Arterioscler Thromb Vasc Biol 2013;33:1706e13. [46] Lopez-Rios L, Perez-Jimenez P, Martinez-Quintana E, et al. Association of Taq 1B CETP polymorphism with insulin and HOMA levels in the population of the Canary Islands. Nutr Metab Cardiovasc Dis 2011;21:18e24. [47] Pihlajamaki J, Karjalainen L, Karhapaa P, et al. G-250A substitution in promoter of hepatic lipase gene is associated with dyslipidemia and insulin resistance in healthy control subjects and in members of families with familial combined hyperlipidemia. Arterioscler Thromb Vasc Biol 2000;20:1789e95. [48] Gomez P, Perez-Jimenez F, Marin C, et al. The -514 C/T polymorphism in the hepatic lipase gene promoter is associated with insulin sensitivity in a healthy young population. J Mol Endocrinol 2005;34:331e8. [49] Stefan N, Schafer S, Machicao F, et al. Liver fat and insulin resistance are independently associated with the -514C>T polymorphism of the hepatic lipase gene. J Clin Endocrinol Metab 2005;90:4238e43. [50] von Eckardstein A. Differential diagnosis of familial high density lipoprotein deficiency syndromes. Atherosclerosis 2006;186:231e9. [51] Vergeer M, Brunham LR, Koetsveld J, et al. Carriers of loss-of-function mutations in ABCA1 display pancreatic beta-cell dysfunction. Diabetes Care 2010;33:869e74. [52] Hegele RA. Phenomics, lipodystrophy, and the metabolic syndrome. Trends Cardiovasc Med 2004;14:133e7. [53] Simha V, Garg A. Lipodystrophy: lessons in lipid and energy metabolism. Curr Opin Lipidol 2006;17:162e9. [54] Hegele RA, Pollex RL. Genetic and physiological insights into the metabolic syndrome. Am J Physiol Regul Integr Comp Physiol 2005;289:R663e9. [55] Teslovich TM, Musunuru K, Smith AV, et al. Biological, clinical and population relevance of 95 loci for blood lipids. Nature 2010;466:707e13. [56] Asselbergs FW, Guo Y, van Iperen EP, et al. Large-scale gene-centric meta-analysis across 32 studies identifies multiple lipid loci. Am J Hum Genet 2012;91:823e38. [57] Wolfs MG, Li N, Fu J, Wijmenga C, van Haeften TW, Hofker MH. Genetic insights through genome wide association studies in type 2 diabetes mellitus will lead to new therapeutics. In: Neri C, editor. Changing views on living organisms. Bentham eBooks; 2013. pp. 210e40. [58] Katan MB. Apolipoprotein E isoforms, serum cholesterol, and cancer. Lancet 1986;1:507e8. [59] Smith GD, Ebrahim S. ‘Mendelian randomization’: can genetic epidemiology contribute to understanding environmental determinants of disease? Int J Epidemiol 2003;32:1e22. [60] Smith GD, Ebrahim S. Mendelian randomization: prospects, potentials, and limitations. Int J Epidemiol 2004;33:30e42. [61] De Silva NM, Freathy RM, Palmer TM, et al. Mendelian randomization studies do not support a role for raised circulating triglyceride levels influencing type 2 diabetes, glucose levels, or insulin resistance. Diabetes 2011;60:1008e18. [62] Qi Q, Liang L, Doria A, Hu FB, Qi L. Genetic predisposition to dyslipidemia and type 2 diabetes risk in two prospective cohorts. Diabetes 2012;61:745e52. [63] Sattar N, Preiss D, Murray HM, et al. Statins and risk of incident diabetes: a collaborative meta-analysis of randomised statin trials. Lancet 2010;375: 735e42.

138

N. Li et al. / Atherosclerosis 233 (2014) 130e138

[64] Baker WL, Talati R, White CM, Coleman CI. Differing effect of statins on insulin sensitivity in non-diabetics: a systematic review and meta-analysis. Diabetes Res Clin Pract 2010;87:98e107. [65] Freeman DJ, Norrie J, Sattar N, et al. Pravastatin and the development of diabetes mellitus: evidence for a protective treatment effect in the West of Scotland Coronary Prevention Study. Circulation 2001;103:357e62. [66] Saito Y. Pitavastatin: an overview. Atheroscler Suppl 2011;12:271e6. [67] Kawai Y, Sato-Ishida R, Motoyama A, Kajinami K. Place of pitavastatin in the statin armamentarium: promising evidence for a role in diabetes mellitus. Drug Des Dev Ther 2011;5:283e97. [68] Tamaki N, Ueno H, Morinaga Y, Shiiya T, Nakazato M. Ezetimibe ameliorates atherosclerotic and inflammatory markers, atherogenic lipid profiles, insulin sensitivity, and liver dysfunction in Japanese patients with hypercholesterolemia. J Atheroscler Thromb 2012;19:532e8. [69] Park H, Shima T, Yamaguchi K, et al. Efficacy of long-term ezetimibe therapy in patients with nonalcoholic fatty liver disease. J Gastroenterol 2011;46: 101e7. [70] Ooi CP, Loke SC. Colesevelam for type 2 diabetes mellitus. Cochrane Database Syst Rev 2012;12:CD009361. [71] Rosenson RS. Colesevelam HCl reduces LDL particle number and increases LDL size in hypercholesterolemia. Atherosclerosis 2006;185:327e30. [72] Goldfine AB. Modulating LDL cholesterol and glucose in patients with type 2 diabetes mellitus: targeting the bile acid pathway. Curr Opin Cardiol 2008;23: 502e11. [73] Smushkin G, Sathananthan M, Piccinini F, et al. The effect of a bile acid sequestrant on glucose metabolism in subjects with type 2 diabetes. Diabetes 2013;62:1094e101. [74] Marina AL, Utzschneider KM, Wright LA, Montgomery BK, Marcovina SM, Kahn SE. Colesevelam improves oral but not intravenous glucose tolerance by a mechanism independent of insulin sensitivity and beta-cell function. Diabetes Care 2012;35:1119e25. [75] Beysen C, Murphy EJ, Deines K, Chan M, Tsang E, Glass A. Effect of bile acid sequestrants on glucose metabolism, hepatic de novo lipogenesis, and

[76]

[77]

[78]

[79]

[80]

[81]

[82]

[83]

[84]

cholesterol and bile acid kinetics in type 2 diabetes: a randomised controlled study. Diabetologia 2012;55:432e42. Jones IR, Swai A, Taylor R, Miller M, Laker MF, Alberti KG. Lowering of plasma glucose concentrations with bezafibrate in patients with moderately controlled NIDDM. Diabetes Care 1990;13:855e63. Tenenbaum A, Motro M, Fisman EZ, Schwammenthal E, Adler Y, Goldenberg I. Peroxisome proliferator-activated receptor ligand bezafibrate for prevention of type 2 diabetes mellitus in patients with coronary artery disease. Circulation 2004;109:2197e202. Tenenbaum A, Fisman EZ, Boyko V, Benderly M, Tanne D, Haim M. Attenuation of progression of insulin resistance in patients with coronary artery disease by bezafibrate. Arch Intern Med 2006;166:737e41. Tenenbaum H, Behar S, Boyko V, Adler Y, Fisman EZ, Tanne D. Long-term effect of bezafibrate on pancreatic beta-cell function and insulin resistance in patients with diabetes. Atherosclerosis 2007;194:265e71. Fabbrini E, Mohammed BS, Korenblat KM, Magkos F, McCrea J, Patterson BW. Effect of fenofibrate and niacin on intrahepatic triglyceride content, very lowdensity lipoprotein kinetics, and insulin action in obese subjects with nonalcoholic fatty liver disease. J Clin Endocrinol Metab 2010;95:2727e35. Bajaj M, Suraamornkul S, Hardies LJ, Glass L, Musi N, DeFronzo RA. Effects of peroxisome proliferator-activated receptor (PPAR)-alpha and PPAR-gamma agonists on glucose and lipid metabolism in patients with type 2 diabetes mellitus. Diabetologia 2007;50:1723e31. Drew BG, Duffy SJ, Formosa MF, et al. High-density lipoprotein modulates glucose metabolism in patients with type 2 diabetes mellitus. Circulation 2009;119:2103e11. Barter PJ, Rye KA, Tardif JC, et al. Effect of torcetrapib on glucose, insulin, and hemoglobin A1c in subjects in the Investigation of Lipid Level Management to Understand its Impact in Atherosclerotic Events (ILLUMINATE) trial. Circulation 2011;124:555e62. Schaub MA, Boyle AP, Kundaje A, Batzoglou S, Snyder M. Linking disease associations with regulatory information in the human genome. Genome Res 2012;22:1748e59.