Tweet
Statin treatment - Clinical implications of pharmacogenomics of
Clinical Implication
The Pharmacogenomics Journal advance online publication 21 March 2006; doi: 10.1038/sj.tpj.6500384
Clinical implications of pharmacogenomics of statin treatment
L M Mangravite<SUP minmax_bound="true">1</SUP>, C F Thorn<SUP minmax_bound="true">2</SUP> and R M Krauss<SUP minmax_bound="true">1</SUP>
Received 1 February 2006; Accepted 6 February 2006; Published online 21 March 2006.
<!-- articlebody start -->
-hydroxy--methylglutaryl Coenzyme A (HMG-CoA) reductase inhibitors, or statins, inhibit endogenous cholesterol production by competitive inhibition of HMG-CoA reductase (HMGCR), the enzyme that catalyzes conversion of HMG-CoA to mevalonate, an early rate-limiting step in cholesterol synthesis. By reducing intracellular cholesterol production, statin treatment results in upregulation of low-density lipoprotein (LDL) receptors, leading to increased plasma clearance of LDL, primarily by the liver. In addition, statins can reduce hepatic secretion of the ApoB-containing lipoproteins, very low-density lipoprotein (VLDL) and LDL. As a result of these effects, statins can reduce plasma levels of atherogenic LDL by as much as 50%. Other effects of potential clinical significance include reductions in plasma triglycerides (TGs), increases in high-density lipoprotein (HDL) cholesterol (HDLC), an indicator of reduced cardiovascular disease (CVD) risk, and reductions in inflammatory markers, notably C-reactive protein (CRP), that have been implicated in the development of CVD.
Statin therapy has been shown in numerous large clinical trials to reduce risk of cardiovascular events by 20?30%, an effect strongly related to the magnitude of LDL cholesterol (LDLC) reduction.<SUP minmax_bound="true">1, 2</SUP> On the basis of these findings, statin treatment in conjunction with lifestyle changes is indicated as first-line therapy for prevention of CVD in individuals who are considered to be at risk.<SUP minmax_bound="true">3</SUP> Adoption of current guidelines for plasma LDL reduction has led to the widespread and increasing use of statins, which are now the most prescribed class of drugs worldwide.
Clinical response to statin-mediated reduction of lipid and lipoprotein parameters is highly variable.<SUP minmax_bound="true">4</SUP> Although statin dosages are often adjusted once individual response to treatment is assessed, nearly a third of statin-treated patients do not meet their lipid-lowering goals.<SUP minmax_bound="true">5</SUP> In addition, adverse drug reactions (ADR), although rare, can be severe. Variability in response to statin therapy results from environmental and non-genetic factors, such as age, gender, diet, smoking status, and physical activity. Just as interindividual variability in plasma lipid and lipoprotein levels is governed by hereditary factors, it stands to reason that statin-response of these same parameters is also related to genetic heterogeneity. In fact, a recent study indicates clear population differences in rosuvastatin sensitivity between subjects of Caucasian, Chinese, Malaysian, and Indian decent all residing in Singapore that could not be accounted for by non-genetic factors.<SUP minmax_bound="true">6</SUP>
This review will summarize studies examining genetic influences on statin efficacy and toxicity, and discuss the potential for this information to guide the optimal clinical use of these compounds.
Top of page Genetic influences on statin efficacy
Genes involved in pharmacokinetic response (Figure 1)
Genetic variations affecting statin pharmacokinetics can alter duration and magnitude of drug exposure, and hence both efficacy and toxicity (Table 1). Efficacy of statin response, measured by either lipid-lowering response or reduction in mortality, is dependent on hepatic rather than systemic statin exposure as these compounds undergo extensive first-pass clearance and the liver is the major site of action.<SUP minmax_bound="true">7</SUP> Although unlikely to affect statin efficacy, genetic variation causing alterations in systemic statin exposure may create susceptibility to adverse drug reactions. Genetic variations affecting hepatic exposure are more likely candidates for altering treatment efficacy.
Table 1 - Pharmacogenetic and pharmacogenomic studies on genes involved in pharmacokinetic handling of statins.
Full table <!-- clearing div -->
There are six statin compounds currently on the market for use as cholesterol-lowering therapies: simvastatin, pravastatin, atorvastatin, lovastatin, fluvastatin, and rosuvastatin. The pharmacokinetic profiles of these compounds vary based on hydrophobicity. The more hydrophilic compounds, pravastatin in particular, require active transport into the liver, are less metabolized by the cytochrome P450 (CYP) family, and exhibit more pronounced active renal excretion; whereas the less hydrophilic compounds are transported by passive diffusion and are better substrates for both CYP enzymes and transporters involved in biliary excretion.<SUP minmax_bound="true">8, 9, 10</SUP> Given the differential involvement of pharmacokinetic genes in the metabolism of statin compounds, variation in these genes may aid in determining treatment choice.
Drug metabolizing enzymes affecting statin therapy
Statins undergo metabolism largely via the CYP3A (lovastatin, atorvastatin, simvastatin) or CYP2C (fluvastatin) families of metabolizing enzymes.<SUP minmax_bound="true">11</SUP> Metabolism of these compounds may also be mediated, in part, by CYP2D6 or several glycosyltransferases (UGT1A1, UGT1A3, UGT2B7).<SUP minmax_bound="true">7, 12</SUP> Pravastatin and rosuvastatin interact minimally with metabolizing enzymes, are largely excreted unchanged, and are less likely to be affected by genetic variation in metabolizing enzymes.
Polymorphisms in genes encoding several of these enzymes have been examined for associations with variability in statin efficacy or systemic exposure. Within the CYP3A family, there are four independent reports of association with lipid lowering response.<SUP minmax_bound="true">13, 14, 15, 16, 17, 18</SUP> None of these observations have survived replication. Within the CYP2C family, the CYP2C9<SUP minmax_bound="true">*</SUP>3 haplotype has been associated with duration of systemic exposure to fluvastatin, as measured by plasma area under the curve (AUC), but not with cholesterol reduction.<SUP minmax_bound="true">19</SUP> Contribution of genetic variation in CYP2D6 to simvastatin efficacy has also been studied with disparate results. An initial study suggested that lipid response to simvastatin treatment was inversely related to enzymatic activity.<SUP minmax_bound="true">20</SUP> This finding was replicated in one of two subsequent studies, both of which enrolled less than 100 subjects.<SUP minmax_bound="true">21, 22, 23</SUP> These data are further complicated by conflicting reports as to the significance of CYP2D6 metabolism in simvastatin clearance.<SUP minmax_bound="true">24, 25, 26</SUP>
Conflicting conclusions from individual studies are largely due to limited sample size as underpowered studies are less likely to yield reproducible results. Two studies have recently appeared that attempted to address this issue: the Pravastatin Inflammation/CRP Evaluation (PRINCE) study, which examined associations of pravastatin efficacy with variation across 10 candidate genes in 1536 subjects; the Atorvastatin Comparative Cholesterol Efficacy and Safety Study (ACCESS), which examined associations between variation in 14 candidate genes and statin efficacy in 3916 subjects on various statins, half of whom were administered atorvastatin.<SUP minmax_bound="true">13, 17</SUP> Both of these studies included CYP3A4 and CYP3A5 as candidate genes, but neither identified significant associations. Whereas pravastatin is minimally metabolized by either CYP3A4 or CYP3A5, both enzymes metabolize atorvastatin. Hence, any influences of these genes on statin efficacy might have been expected to be detected in the ACCESS cohort.
Drug transporter proteins affecting statin therapy
Transporter proteins appear to be involved in hepatobiliary elimination of all statins as well as in absorption, distribution, and renal excretion of the more hydrophilic compounds. Although intestinal absorption of statins, all of which are dosed orally, is poorly studied, it is assumed to occur mainly by passive diffusion. There is some evidence for interaction with proton-dependent active transporters: specifically, SLCO2B1 in the case of pravastatin and SLC15A1 in the case of fluvastatin.<SUP minmax_bound="true">27, 28</SUP> SLCO2B1 is also implicated in hepatic uptake of pravastatin and has been examined for genetic contributions to interindividual variation in response to pravastatin. The main variant identified as having potential for pharmacogenetic response, marked by haplotype SLCO2B1<SUP minmax_bound="true">*</SUP>3, encodes a transporter of reduced function in vitro, but has not been associated with altered systemic drug exposure in vivo.<SUP minmax_bound="true">29, 30</SUP> Pravastatin also appears to be dependent on active transport for renal elimination, through SLC22A8 and SLC22A6.
Active transport is also important for hepatic uptake of statins, particularly pravastatin, and appears to be dependent on SLCO1B1 and to a lesser degree on SLCO2B1 and SLCO1B3. The most common genetic variants of SLCO1B1 (A388G, T521C) are represented alone or together as haplotypes <SUP minmax_bound="true">*</SUP>1A (neither), <SUP minmax_bound="true">*</SUP>1B (A388G), <SUP minmax_bound="true">*</SUP>5 (T521C), and <SUP minmax_bound="true">*</SUP>15 (A388G, T521C). Several single-dose pravastatin studies indicated that the <SUP minmax_bound="true">*</SUP>1B haplotype was associated with decreased plasma AUC, implying accelerated hepatocellular uptake, whereas the <SUP minmax_bound="true">*</SUP>5 and <SUP minmax_bound="true">*</SUP>15 haplotypes were associated with increased plasma AUC, indicating delayed hepatocellular uptake.<SUP minmax_bound="true">29, 31, 32</SUP> An additional haplotype, SLCO1B1<SUP minmax_bound="true">*</SUP>17, formed by the presence of a promoter variant (G-11187A), was also associated with elevated plasma AUC and reduced intracellular cholesterol synthesis.<SUP minmax_bound="true">33</SUP> SLCO1B1<SUP minmax_bound="true">*</SUP>5 carriers were reported to have attenuated plasma total cholesterol (TC) reduction in comparison to non-carriers.<SUP minmax_bound="true">34</SUP>
Hepatobiliary excretion of statins is mediated by ABCC2 and ABCB1 as well as ABCG2 and ABCB11, all of which belong to a family of transporters known to interact with lipophilic xenobiotics.<SUP minmax_bound="true">35, 36, 37, 38, 39, 40</SUP> Variations in these transporters could alter duration of hepatic exposure, and therefore exposure to sites of action and to metabolizing enzymes. Variation in ABCC2, likely the largest contributor to biliary excretion of statins, is known to exhibit variation; however, this has been poorly studied in the context of statin transport. The only report notes that variation does not appear to alter plasma pravastatin AUC,<SUP minmax_bound="true">29</SUP> a finding consistent with the fact that this transporter is involved in hepatic export rather than import. ABCB1 encodes P-glycoprotein, which has been implicated in the transport of many drugs. A number of non-synonymous ABCB1variants responsible for altered protein function, including C1236T, C3435T and G2677T/A, have been tested for associations with the lipid-lowering efficacy of statins, without definitive results.<SUP minmax_bound="true">14, 41, 42</SUP> ABCC2 and ABCB1 are both involved in efflux of a wide variety of commonly prescribed drugs and conceivably functional variations in these transporters in conjunction with co-administration of interacting therapies could increase systemic drug exposures and risk for adverse drug reactions.<SUP minmax_bound="true">11</SUP>
Genes involved in pharmacodynamic response (Figure 2)
Genetic variations in any of the multiple genes involved in cholesterol metabolism, lipoprotein secretion, or lipoprotein clearance have the potential to affect the magnitude of plasma lipid-lowering response to statins and ultimately the extent to which these responses lead to reduction in CVD risks (Table 2). Additionally, statins are known to have pleiotropic non-lipid-lowering effects that may influence CVD events via immune and inflammatory pathways or altered endothelial function.<SUP minmax_bound="true">43</SUP> Among likely mechanisms of interaction between statins and these pathways are alterations in production of cholesterol precursor molecules common to both lipid and non-lipid pathways. As production of these molecules is downstream to HMGCR-catalyzed mevalonate production, they may also be depleted by statin treatment.
Table 2 - Pharmacogenetic and pharmacogenomic studies on genes involved in pharmacodynamic response to statins.
Full table <!-- clearing div -->
Genetic variation in HMGCR, the direct target of statin therapy, is surprisingly understudied. The PRINCE study has reported two non-coding HMGCR variants in tight linkage disequilibrium (SNPs12 and 29) that associated with magnitude of LDLC response. Carriers of the haplotype constituted by these SNPs displayed smaller reductions in LDLC than non-carriers.<SUP minmax_bound="true">13</SUP> These were the only two SNPs of 148 found to meet the rigorous conditions for significance defined by this study, which included adjustments for multiple testing, similarity of effects in both the whole population and the Caucasian subpopulation (representing 88.7% of the subjects), and lack of association with baseline lipid levels. These criteria were intended to define variants strictly involved in pharmacogenetic response as well as to reduce likelihood of Type I error. Even so, association between these HMGCR SNPs and LDLC response failed to replicate in the ACCESS cohort.<SUP minmax_bound="true">17</SUP> Whereas these initial findings may not be indicative of a true pharmacogenetic effect, there are other possible explanations for these discordant results. In particular, as neither SNP is predicted to alter function or expression of HMGCR, these SNPs may be in incomplete linkage disequilibrium (LD) with a causal variant. The extent of LD, as well as other background genetic or environmental influences, may differ in the two study populations. It may also be that there are differing pharmacogenetic effects of the HMGCR variant for pravastatin and atorvastatin.
Of the many genes involved in cholesterol biosynthesis and catabolism, polymorphisms in only two others have been examined for association with statin response. Squalene synthase (FDFT1) encodes a protein downstream to HMGCR in the cholesterol biosynthesis pathway. Variation in FDFT1 could also play a role in pleiotropic response to statin treatment, as it is upstream to production of molecules that may mediate these effects. Variation in this gene was not associated with lipid response to statin therapy.<SUP minmax_bound="true">13, 17</SUP> A promoter variant in CYP7A1, which encodes an enzyme that has a critical function in cholesterol oxidation and excretion, has been examined for association with diminished reduction in atorvastatin-induced LDLC in two studies, with significant association seen only in one.<SUP minmax_bound="true">17, 44</SUP>
Many genes involved in lipoprotein structure, secretion, and catabolism have been examined for influences on statin response. The most frequently studied of these are LDL receptor (LDLR), cholesterol ester transfer protein (CETP), and apolipoprotein E (ApoE). LDLR is a primary candidate for pharmacogenetic testing as it is directly implicated in the mechanism of statin-mediated LDL reduction. In addition, rare loss of function mutations in the gene encoding LDLR cause familial hypercholesterolemia (FH), a condition marked by strikingly elevated plasma cholesterol levels and increased risk of mortality. Statin-treated FH patients show variable responses to treatment in an LDLR mutation-dependent manner. Carriers of loss of function mutations appear more responsive than null function carriers.<SUP minmax_bound="true">45, 46, 47, 48, 49</SUP> More common, less dysfunctional LDLR genetic variants could alter response in non-FH hypercholesterolemic individuals. Several small studies suggest that this is true, although none of these associations were replicated in larger populations such as PRINCE or ACCESS.<SUP minmax_bound="true">13, 17, 50, 51</SUP>
Genes involved in hepatic regulation of LDLR expression are also hypothesized to contribute to genetic variation in statin response. LDLR expression is highly regulated by many intracellular pathways, most significantly through transcriptional regulation by the sterol regulatory element binding protein (SREBP) family of transcription factors. These are sequestered in the endoplasmic reticulum under normal conditions and, when signaled by sterol depletion, are escorted by SREBP cleavage activating protein (SCAP) into the Golgi for proteolytic activation. Variations in SREBP1, SREBP2, and SCAP have been examined for genetic associations to statin response. SREBP1 promoter variant G-36D predicted the magnitude of statin-induced ApoAI elevation in one of two studies, whereas a polymorphism in SCAP has been reported to influence response of LDLC, TC, and TG.<SUP minmax_bound="true">52, 53</SUP> Association of SCAP variation with response of multiple lipid parameters suggests validity, but the SCAP variant associations have not replicated.<SUP minmax_bound="true">53, 54</SUP> Recently, a novel protease, PCSK9, has been implicated in LDLR protein degradation and variation in this gene was associated with basal plasma LDLC.<SUP minmax_bound="true">55, 56, 57</SUP> PCSK9 gain-of-function mutations causing increased degradation of LDLR and reduced clearance of plasma LDL were associated with higher pretreatment LDLC levels and attenuated statin-mediated reduction of LDLC.<SUP minmax_bound="true">55, 56</SUP> This newly discovered element of the lipoprotein regulatory system may be an important contributor to statin response.
The most closely examined gene connected with lipoprotein structure and metabolism is that encoding ApoE, a protein that resides on triglyceride-rich atherogenic lipoproteins as well as HDL, and interacts with LDLR to promote hepatic particle clearance. Three major variants of this form have been identified;
3, which encodes wild type protein, 2 and 4, both of which contain one non-synonymous variant causing altered protein function. The presence of variant isoforms causes altered particle clearance and is associated with hyperlipidemic states. Many studies have examined the contributions of these isoforms to statin-mediated response. The consensus is that 4 carriers appear to have attenuated lipid-lowering response, and 2 carriers have enhanced response.<SUP minmax_bound="true">13, 17, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70</SUP> Few studies have examined the effects of variations in genes encoding other apolipoproteins. There are reports that polymorphisms in APOB, the gene encoding the major structural protein in both LDL and triglyceride-rich lipoproteins, may alter LDLC response; whereas variation in the APOAI gene, encoding the major structural protein in HDL, may alter HDLC response.<SUP minmax_bound="true">13, 17, 50, 62, 66, 71, 72, 73</SUP>
Modulation of lipoprotein lipid content is carried out by a series of proteins that act through mechanisms of inter-lipoprotein lipid transfers and lipolysis. CETP catalyzes the transfer of lipids between HDL and triglyceride-rich lipoproteins. Variation in this gene, particularly a TaqIB variant located within intron 1, has been reported to be associated with plasma HDL concentration and CVD risk under basal conditions.<SUP minmax_bound="true">74, 75</SUP> Many studies have examined the influence of the TaqIB variant on statin response. Initial reports suggesting that TaqIB is associated with variation in HDLC response and CVD disease end points have failed to replicate.<SUP minmax_bound="true">13, 76, 77, 78</SUP> Boekholdt et al.<SUP minmax_bound="true">75</SUP> recently performed a meta-analysis, using data from 13 500 subjects across 10 trials but could not confirm this result. Although this variant may have a minor effect on response to statin therapy, it appears that the magnitude of the contribution is too small to easily distinguish from other sources of variation. Other lipid-modulating genes that have been studied include microsomal triglyceride transfer protein (MTP), which also encodes for a lipid transfer protein, and the lipase genes, lipoprotein lipase (LPL) and hepatic lipase (LIPC). Single studies have reported that a promoter variant in MTP is associated with triglyceride response and recurring risk of coronary events, an LPL variant may alter TC or HDLC response, and an LIPC variant may alter HDLC response.<SUP minmax_bound="true">79, 80, 81, 82, 83, 84, 85</SUP>
Removal of cholesterol from peripheral cells requires active transport and is mediated by ABCA1. Preliminary pharmacogenetic analysis suggests that an ABCA1 haplotype may attenuate ApoAI response to statin treatment.<SUP minmax_bound="true">86</SUP>
ABCG5 and ABCG8 are transporters that act within the enterocyte to limit intestinal absorption, and within the hepatocyte to increase biliary excretion.<SUP minmax_bound="true">87, 88</SUP> Rare, loss of function mutations in these genes lead to sitosterolemia, a disorder characterized by hypercholesterolemia caused by increased intestinal cholesterol absorption and decreased biliary secretion.<SUP minmax_bound="true">89</SUP> Intestinal absorption of cholesterol is tightly regulated in conjunction with hepatic cholesterol synthesis to maintain homeostatic levels of systemic cholesterol.<SUP minmax_bound="true">90</SUP> LDLC resistance to statin therapy may be related to a compensatory increase in cholesterol absorption via upregulation of ABCG5 and ABCG8.<SUP minmax_bound="true">91, 92, 93</SUP> As such, these are major candidates for pharmacogenetic examination. To date, there is suggestive evidence that an ABCG8 variant is associated with LDLC response to statin treatment.<SUP minmax_bound="true">13, 17, 92</SUP>
Pathways governing the pleiotropic or non-lipid response to statin therapy are under active investigation. There is evidence that these pathways contribute to statin-induced reductions in coronary events and mortality but their importance in this regard is not well established. As these pathways are incompletely characterized, it is difficult to identify candidate genes that may have pharmacogenetic influence. Nevertheless, a handful of studies have examined variation in putative response genes. As the mechanisms involved are independent of lipid-lowering effects, response in these studies tends to be defined by number of post-treatment coronary events or degree of atherosclerotic lesion progression. Replicate studies have been reported for only three genes: ACE, IGTB3, and TLR4. The gene encoding angiotensin converting enzyme (ACE) plays a role in regulation of blood pressure. An insertion/deletion variant within the ACE gene has been reported to predict risk of recurring coronary events following statin treatment in several studies.<SUP minmax_bound="true">94, 95, 96</SUP> The IGTB3 gene encodes the platelet-specific fibrinogen receptor and variation in this gene has been associated under basal conditions with elevated risk of coronary disease.<SUP minmax_bound="true">94, 97</SUP> This increase in risk was reported to be abolished by statin treatment.<SUP minmax_bound="true">94, 97</SUP> The TLR4 gene encodes for toll-like receptor 4, which is involved in innate immunity. Carriers of a non-synonymous TLR4 variant (D299G) demonstrated reduced risk of post-treatment coronary events.<SUP minmax_bound="true">98, 99</SUP> None of the mechanisms for these associations is known. In short, studies regarding the role of variation in genes involved in non-lipid-lowering response to statins have not been definitive. A clearer understanding of the components of these pathways is required before a suitable list of candidate genes can be examined.
Although variation in genes involved in cholesterol biosynthesis or lipoprotein production and clearance are established contributors to variation in basal concentrations of plasma lipids, studies to date have not definitely demonstrated a link to variation in statin-mediated lipid response. The lack of success of studies aimed at discovering the genetic variation contributing to therapeutic response is due partly to methodological limitations, most notably lack of sufficient statistical power, and partly to the nature of the goal. Statin-mediated lipid reduction is a complicated system with many components. Individual examination of isolated genetic variants is unlikely to provide useful answers. Recent studies have moved to a multiple candidate gene strategy, but these have not yet examined the influence of interactions between these variations. This approach, as well as genome-wide association or linkage studies using larger populations, may prove more productive than previous studies.
Top of page Genetic influences on statin-mediated adverse drug reactions
A particularly promising application of pharmacogenetics in statin therapy is the use of genetic pre-screening to identify subjects susceptible to adverse drug reactions. Although statin therapy is relatively well tolerated, with mild dyspepsia and gastrointestinal discomfort accounting for the majority of reported adverse reactions, more severe reactions can occur, notably severe hepatic dysfunction (
0.7% frequency) and skeletal myopathy (0.1% frequency). These have been reported with all statins to varying degrees.<SUP minmax_bound="true">11</SUP> Although these events are rare, given the exceptional number of patients prescribed these drugs, the occurrences in real terms are considerable and serious enough to cause public concern.<SUP minmax_bound="true">100</SUP> Muscle biopsies from patients indicated that myopathy is defined on the cellular level by mitochondrial dysfunction, increased lipid storage, and ragged red muscle fibers.<SUP minmax_bound="true">101</SUP> Approximately 2?5% of myopathies lead to rhabdomyolosis, characterized by muscle destruction and myoglobinuria.<SUP minmax_bound="true">101</SUP> As myopathy is linked to elevated creatine kinase levels, monitoring of creatine kinase concentrations during statin treatment is recommended to identify cases before symptoms become severe.<SUP minmax_bound="true">102</SUP> Notably, not all cases of statin-induced myopathy are accompanied by elevated creatine kinase.<SUP minmax_bound="true">103</SUP>
Risk and severity of myopathy with stain treatment, unlike statin efficacy, is related to magnitude and duration of systemic exposure.<SUP minmax_bound="true">102</SUP> Thus, the risk for this complication is more susceptible to any source of variation in pharmacokinetic parameters. Systemic exposure is affected by many non-genetic factors, including age, renal and hepatic function, smoking status, and concomitant medications. It is also likely to be affected by polymorphisms in the genes involved in pharmacokinetic handling of statins. Many of the studies discussed above examined the effect of variation in these genes on plasma AUC. Although this is not an appropriate predictor of efficacy, it may be a useful predictor of toxicity. Additionally, some of the transporters noted above, SLC22A6/8 and ABCB1, in particular, may be involved in transport of statins into skeletal muscle cells.<SUP minmax_bound="true">14, 17, 104</SUP> Current reports suggest that the SLC16 family of transporters, responsible for lactic acid transport across plasma membrane into muscle, also interacts with statins.<SUP minmax_bound="true">105, 106</SUP> Inhibition of SLC16A4, which is primarily expressed in skeletal but not cardiac myocytes, paralleling the site of clinical myopathy, is reported to attenuate statin-induced myopathy in vitro.<SUP minmax_bound="true">106, 107</SUP> This may be a good candidate gene for statin toxicity studies.
Pharmacogenetic analysis of statin myopathy is limited partly because the physiological mechanisms underlying statin-induced myopathy are not well understood. Myopathy may be linked to mitochondrial dysfunction, particularly within Complex I of the respiratory chain.<SUP minmax_bound="true">108</SUP> Initially thought to result from intracellular cholesterol depletion and alterations of membrane fluidity, myopathy is currently thought to be independent of cholesterol depletion and has been suggested to depend on depletion of Coenzyme Q, also a downstream derivative of mevalonate.<SUP minmax_bound="true">109</SUP> Coenzyme Q10 is a member of the mitochondrial respiratory chain as well as having involvement in maintaining cellular health.<SUP minmax_bound="true">110, 111</SUP> Both plasma and muscular concentrations of Coenzyme Q10 are depleted by statin therapy.<SUP minmax_bound="true">112, 113</SUP> Although Coenzyme Q10 supplements during statin therapy have been advocated for relieving muscle discomfort and reducing risk for myopathy, there is no conclusive evidence for their effectiveness.<SUP minmax_bound="true">114</SUP> Alternatively, statins may cause myopathy by impairing intracellular ion pumps, altering Na+ and Ca2+ concentrations, and impairing cell conductance or inducing apoptosis.<SUP minmax_bound="true">115, 116</SUP>
Pharmacogenetic analysis of statin-induced myopathy is also limited by availability of cases. With a frequency of 0.1%, even the largest of the statin trials include only a handful of serious myopathy cases, making genetic association studies very difficult to power. Analyses to date are based on case studies or examination of a small number of subjects. Variation in CYP2D6 has been associated with decreased adherence to simvastatin therapy, although the metabolic significance of this enzyme in simvastatin clearance is under debate.<SUP minmax_bound="true">23, 25</SUP> The presence of the4 allele of ApoE is associated with decreased adherence.<SUP minmax_bound="true">67</SUP> A case of rhabdomyolysis-induced death from cerivastatin exposure may have been caused by a loss of function mutation in CYP2C8 resulting in increased systemic exposure.<SUP minmax_bound="true">117</SUP> The SLCO1B1<SUP minmax_bound="true">*</SUP>15 haplotype, causing decreased hepatic uptake of pravastatin, is over-represented in subjects presenting with rhabdomyolysis.<SUP minmax_bound="true">118</SUP>
Novel methods of recruitment will be necessary to compile populations better served to address the question of genetic variation in statin-induced ADR. Recently Wilke et al.<SUP minmax_bound="true">119</SUP> examined 68 cases of statin-induced myopathy within a retrospective case?control study of patients identified from the Marshfield Clinic patient population for association between CYP3A genotype and risk or severity of myopathy, as measured by serum CK level. Results indicated that severity but not risk was associated with the presence of two copies of the CYP3A5<SUP minmax_bound="true">*</SUP>3 haplotype. Ruano et al. examined 19 SNPs from 10 candidate genes involved in vascular homeostasis for association with statin-induced elevated serum CK levels and reported that polymorphisms in the gene for angiotensin II receptor 1(AGTR1), a multifunctional protein, and the gene encoding nitric oxide synthase 3 (NOS3), an enzyme involved in smooth muscle contraction and platelet aggregation, were associated with serum CK.<SUP minmax_bound="true">120</SUP> These two studies represent the first major examinations of genetic variation in statin-induced myopathy.
Top of page Summary of clinical implications
Although, statin pharmacogenetics is still in its infancy, the consensus based on current evidence is that genetic testing to assist in selecting a particular statin and/or dosing regimen is not clinically warranted. There are several reasons for this opinion. Statin therapy is so well tolerated in the majority of patients that the need to use additional testing to avoid deleterious side effects is relatively small. Moreover patients should be monitored for signs of hepatic or muscle toxicity to avoid more severe adverse reactions. In addition, patients on statin therapy should have lipid levels monitored so that submaximal response can be improved by increased dosage. Another major concern is that nearly all studies of pharmacogenetics of statin response reported to date have been underpowered and replication of positive results is lacking. Finally, most of these studies have examined associations of individual genetic polymorphisms with treatment response. Although genetic differences may prove a significant source of response variability, this is unlikely to be due to any single gene. Rather, the compound effects of multiple genetic variants are more likely to be responsible. Most studies have examined a small number of variants in a single or small number of genes ? often one variant in one gene ? and have reported the resultant outcomes based on small patient populations.
There are, however, other considerations that support the potential value of statin pharmacogenetic information. Identification of genes and genetic variants that influence statin responsiveness holds promise for identifying molecular components of physiological lipid and inflammatory pathways that mediate statin effects. However, the most important clinical benefits of statin pharmacogenetic knowledge would be based on identifying a set of genotypes that aid in predicting the outcomes of statin treatment in terms of reduced risk for cardiovascular events, coupled with reduced risk for serious ADRs. As studies to date have documented at most a 30% statin-mediated reduction in risk for CVD events and stroke, it would be valuable to have information on potential genetic influences on the likelihood of a beneficial outcome. However, a meaningful analysis of the genetic determinants responsible for statin efficacy and toxicity will require more comprehensive analysis of genome-wide variations in multiple large-scale populations. Until such information is available for either efficacy or safety, there is little reason to consider the use of genetic testing for guiding statin treatment.
Top of page Notes
Duality of interest
None declared.
The Pharmacogenomics Journal advance online publication 21 March 2006; doi: 10.1038/sj.tpj.6500384
Clinical implications of pharmacogenomics of statin treatment
L M Mangravite<SUP minmax_bound="true">1</SUP>, C F Thorn<SUP minmax_bound="true">2</SUP> and R M Krauss<SUP minmax_bound="true">1</SUP>
- <LI id=aff1 minmax_bound="true"><SUP minmax_bound="true">1</SUP>Department of Atherosclerosis Research, Children's Hospital Oakland Research Institute, Oakland, CA, USA
- <SUP minmax_bound="true">2</SUP>Department of Genetics, Stanford University School of Medicine, Stanford, CA, USA
Received 1 February 2006; Accepted 6 February 2006; Published online 21 March 2006.
<!-- articlebody start -->
-hydroxy--methylglutaryl Coenzyme A (HMG-CoA) reductase inhibitors, or statins, inhibit endogenous cholesterol production by competitive inhibition of HMG-CoA reductase (HMGCR), the enzyme that catalyzes conversion of HMG-CoA to mevalonate, an early rate-limiting step in cholesterol synthesis. By reducing intracellular cholesterol production, statin treatment results in upregulation of low-density lipoprotein (LDL) receptors, leading to increased plasma clearance of LDL, primarily by the liver. In addition, statins can reduce hepatic secretion of the ApoB-containing lipoproteins, very low-density lipoprotein (VLDL) and LDL. As a result of these effects, statins can reduce plasma levels of atherogenic LDL by as much as 50%. Other effects of potential clinical significance include reductions in plasma triglycerides (TGs), increases in high-density lipoprotein (HDL) cholesterol (HDLC), an indicator of reduced cardiovascular disease (CVD) risk, and reductions in inflammatory markers, notably C-reactive protein (CRP), that have been implicated in the development of CVD.Statin therapy has been shown in numerous large clinical trials to reduce risk of cardiovascular events by 20?30%, an effect strongly related to the magnitude of LDL cholesterol (LDLC) reduction.<SUP minmax_bound="true">1, 2</SUP> On the basis of these findings, statin treatment in conjunction with lifestyle changes is indicated as first-line therapy for prevention of CVD in individuals who are considered to be at risk.<SUP minmax_bound="true">3</SUP> Adoption of current guidelines for plasma LDL reduction has led to the widespread and increasing use of statins, which are now the most prescribed class of drugs worldwide.
Clinical response to statin-mediated reduction of lipid and lipoprotein parameters is highly variable.<SUP minmax_bound="true">4</SUP> Although statin dosages are often adjusted once individual response to treatment is assessed, nearly a third of statin-treated patients do not meet their lipid-lowering goals.<SUP minmax_bound="true">5</SUP> In addition, adverse drug reactions (ADR), although rare, can be severe. Variability in response to statin therapy results from environmental and non-genetic factors, such as age, gender, diet, smoking status, and physical activity. Just as interindividual variability in plasma lipid and lipoprotein levels is governed by hereditary factors, it stands to reason that statin-response of these same parameters is also related to genetic heterogeneity. In fact, a recent study indicates clear population differences in rosuvastatin sensitivity between subjects of Caucasian, Chinese, Malaysian, and Indian decent all residing in Singapore that could not be accounted for by non-genetic factors.<SUP minmax_bound="true">6</SUP>
This review will summarize studies examining genetic influences on statin efficacy and toxicity, and discuss the potential for this information to guide the optimal clinical use of these compounds.
Top of page Genetic influences on statin efficacy
Genes involved in pharmacokinetic response (Figure 1)
Genetic variations affecting statin pharmacokinetics can alter duration and magnitude of drug exposure, and hence both efficacy and toxicity (Table 1). Efficacy of statin response, measured by either lipid-lowering response or reduction in mortality, is dependent on hepatic rather than systemic statin exposure as these compounds undergo extensive first-pass clearance and the liver is the major site of action.<SUP minmax_bound="true">7</SUP> Although unlikely to affect statin efficacy, genetic variation causing alterations in systemic statin exposure may create susceptibility to adverse drug reactions. Genetic variations affecting hepatic exposure are more likely candidates for altering treatment efficacy.
Table 1 - Pharmacogenetic and pharmacogenomic studies on genes involved in pharmacokinetic handling of statins.
Full table <!-- clearing div -->There are six statin compounds currently on the market for use as cholesterol-lowering therapies: simvastatin, pravastatin, atorvastatin, lovastatin, fluvastatin, and rosuvastatin. The pharmacokinetic profiles of these compounds vary based on hydrophobicity. The more hydrophilic compounds, pravastatin in particular, require active transport into the liver, are less metabolized by the cytochrome P450 (CYP) family, and exhibit more pronounced active renal excretion; whereas the less hydrophilic compounds are transported by passive diffusion and are better substrates for both CYP enzymes and transporters involved in biliary excretion.<SUP minmax_bound="true">8, 9, 10</SUP> Given the differential involvement of pharmacokinetic genes in the metabolism of statin compounds, variation in these genes may aid in determining treatment choice.
Drug metabolizing enzymes affecting statin therapy
Statins undergo metabolism largely via the CYP3A (lovastatin, atorvastatin, simvastatin) or CYP2C (fluvastatin) families of metabolizing enzymes.<SUP minmax_bound="true">11</SUP> Metabolism of these compounds may also be mediated, in part, by CYP2D6 or several glycosyltransferases (UGT1A1, UGT1A3, UGT2B7).<SUP minmax_bound="true">7, 12</SUP> Pravastatin and rosuvastatin interact minimally with metabolizing enzymes, are largely excreted unchanged, and are less likely to be affected by genetic variation in metabolizing enzymes.
Polymorphisms in genes encoding several of these enzymes have been examined for associations with variability in statin efficacy or systemic exposure. Within the CYP3A family, there are four independent reports of association with lipid lowering response.<SUP minmax_bound="true">13, 14, 15, 16, 17, 18</SUP> None of these observations have survived replication. Within the CYP2C family, the CYP2C9<SUP minmax_bound="true">*</SUP>3 haplotype has been associated with duration of systemic exposure to fluvastatin, as measured by plasma area under the curve (AUC), but not with cholesterol reduction.<SUP minmax_bound="true">19</SUP> Contribution of genetic variation in CYP2D6 to simvastatin efficacy has also been studied with disparate results. An initial study suggested that lipid response to simvastatin treatment was inversely related to enzymatic activity.<SUP minmax_bound="true">20</SUP> This finding was replicated in one of two subsequent studies, both of which enrolled less than 100 subjects.<SUP minmax_bound="true">21, 22, 23</SUP> These data are further complicated by conflicting reports as to the significance of CYP2D6 metabolism in simvastatin clearance.<SUP minmax_bound="true">24, 25, 26</SUP>
Conflicting conclusions from individual studies are largely due to limited sample size as underpowered studies are less likely to yield reproducible results. Two studies have recently appeared that attempted to address this issue: the Pravastatin Inflammation/CRP Evaluation (PRINCE) study, which examined associations of pravastatin efficacy with variation across 10 candidate genes in 1536 subjects; the Atorvastatin Comparative Cholesterol Efficacy and Safety Study (ACCESS), which examined associations between variation in 14 candidate genes and statin efficacy in 3916 subjects on various statins, half of whom were administered atorvastatin.<SUP minmax_bound="true">13, 17</SUP> Both of these studies included CYP3A4 and CYP3A5 as candidate genes, but neither identified significant associations. Whereas pravastatin is minimally metabolized by either CYP3A4 or CYP3A5, both enzymes metabolize atorvastatin. Hence, any influences of these genes on statin efficacy might have been expected to be detected in the ACCESS cohort.
Drug transporter proteins affecting statin therapy
Transporter proteins appear to be involved in hepatobiliary elimination of all statins as well as in absorption, distribution, and renal excretion of the more hydrophilic compounds. Although intestinal absorption of statins, all of which are dosed orally, is poorly studied, it is assumed to occur mainly by passive diffusion. There is some evidence for interaction with proton-dependent active transporters: specifically, SLCO2B1 in the case of pravastatin and SLC15A1 in the case of fluvastatin.<SUP minmax_bound="true">27, 28</SUP> SLCO2B1 is also implicated in hepatic uptake of pravastatin and has been examined for genetic contributions to interindividual variation in response to pravastatin. The main variant identified as having potential for pharmacogenetic response, marked by haplotype SLCO2B1<SUP minmax_bound="true">*</SUP>3, encodes a transporter of reduced function in vitro, but has not been associated with altered systemic drug exposure in vivo.<SUP minmax_bound="true">29, 30</SUP> Pravastatin also appears to be dependent on active transport for renal elimination, through SLC22A8 and SLC22A6.
Active transport is also important for hepatic uptake of statins, particularly pravastatin, and appears to be dependent on SLCO1B1 and to a lesser degree on SLCO2B1 and SLCO1B3. The most common genetic variants of SLCO1B1 (A388G, T521C) are represented alone or together as haplotypes <SUP minmax_bound="true">*</SUP>1A (neither), <SUP minmax_bound="true">*</SUP>1B (A388G), <SUP minmax_bound="true">*</SUP>5 (T521C), and <SUP minmax_bound="true">*</SUP>15 (A388G, T521C). Several single-dose pravastatin studies indicated that the <SUP minmax_bound="true">*</SUP>1B haplotype was associated with decreased plasma AUC, implying accelerated hepatocellular uptake, whereas the <SUP minmax_bound="true">*</SUP>5 and <SUP minmax_bound="true">*</SUP>15 haplotypes were associated with increased plasma AUC, indicating delayed hepatocellular uptake.<SUP minmax_bound="true">29, 31, 32</SUP> An additional haplotype, SLCO1B1<SUP minmax_bound="true">*</SUP>17, formed by the presence of a promoter variant (G-11187A), was also associated with elevated plasma AUC and reduced intracellular cholesterol synthesis.<SUP minmax_bound="true">33</SUP> SLCO1B1<SUP minmax_bound="true">*</SUP>5 carriers were reported to have attenuated plasma total cholesterol (TC) reduction in comparison to non-carriers.<SUP minmax_bound="true">34</SUP>
Hepatobiliary excretion of statins is mediated by ABCC2 and ABCB1 as well as ABCG2 and ABCB11, all of which belong to a family of transporters known to interact with lipophilic xenobiotics.<SUP minmax_bound="true">35, 36, 37, 38, 39, 40</SUP> Variations in these transporters could alter duration of hepatic exposure, and therefore exposure to sites of action and to metabolizing enzymes. Variation in ABCC2, likely the largest contributor to biliary excretion of statins, is known to exhibit variation; however, this has been poorly studied in the context of statin transport. The only report notes that variation does not appear to alter plasma pravastatin AUC,<SUP minmax_bound="true">29</SUP> a finding consistent with the fact that this transporter is involved in hepatic export rather than import. ABCB1 encodes P-glycoprotein, which has been implicated in the transport of many drugs. A number of non-synonymous ABCB1variants responsible for altered protein function, including C1236T, C3435T and G2677T/A, have been tested for associations with the lipid-lowering efficacy of statins, without definitive results.<SUP minmax_bound="true">14, 41, 42</SUP> ABCC2 and ABCB1 are both involved in efflux of a wide variety of commonly prescribed drugs and conceivably functional variations in these transporters in conjunction with co-administration of interacting therapies could increase systemic drug exposures and risk for adverse drug reactions.<SUP minmax_bound="true">11</SUP>
Genes involved in pharmacodynamic response (Figure 2)
Genetic variations in any of the multiple genes involved in cholesterol metabolism, lipoprotein secretion, or lipoprotein clearance have the potential to affect the magnitude of plasma lipid-lowering response to statins and ultimately the extent to which these responses lead to reduction in CVD risks (Table 2). Additionally, statins are known to have pleiotropic non-lipid-lowering effects that may influence CVD events via immune and inflammatory pathways or altered endothelial function.<SUP minmax_bound="true">43</SUP> Among likely mechanisms of interaction between statins and these pathways are alterations in production of cholesterol precursor molecules common to both lipid and non-lipid pathways. As production of these molecules is downstream to HMGCR-catalyzed mevalonate production, they may also be depleted by statin treatment.
Table 2 - Pharmacogenetic and pharmacogenomic studies on genes involved in pharmacodynamic response to statins.
Full table <!-- clearing div -->Genetic variation in HMGCR, the direct target of statin therapy, is surprisingly understudied. The PRINCE study has reported two non-coding HMGCR variants in tight linkage disequilibrium (SNPs12 and 29) that associated with magnitude of LDLC response. Carriers of the haplotype constituted by these SNPs displayed smaller reductions in LDLC than non-carriers.<SUP minmax_bound="true">13</SUP> These were the only two SNPs of 148 found to meet the rigorous conditions for significance defined by this study, which included adjustments for multiple testing, similarity of effects in both the whole population and the Caucasian subpopulation (representing 88.7% of the subjects), and lack of association with baseline lipid levels. These criteria were intended to define variants strictly involved in pharmacogenetic response as well as to reduce likelihood of Type I error. Even so, association between these HMGCR SNPs and LDLC response failed to replicate in the ACCESS cohort.<SUP minmax_bound="true">17</SUP> Whereas these initial findings may not be indicative of a true pharmacogenetic effect, there are other possible explanations for these discordant results. In particular, as neither SNP is predicted to alter function or expression of HMGCR, these SNPs may be in incomplete linkage disequilibrium (LD) with a causal variant. The extent of LD, as well as other background genetic or environmental influences, may differ in the two study populations. It may also be that there are differing pharmacogenetic effects of the HMGCR variant for pravastatin and atorvastatin.
Of the many genes involved in cholesterol biosynthesis and catabolism, polymorphisms in only two others have been examined for association with statin response. Squalene synthase (FDFT1) encodes a protein downstream to HMGCR in the cholesterol biosynthesis pathway. Variation in FDFT1 could also play a role in pleiotropic response to statin treatment, as it is upstream to production of molecules that may mediate these effects. Variation in this gene was not associated with lipid response to statin therapy.<SUP minmax_bound="true">13, 17</SUP> A promoter variant in CYP7A1, which encodes an enzyme that has a critical function in cholesterol oxidation and excretion, has been examined for association with diminished reduction in atorvastatin-induced LDLC in two studies, with significant association seen only in one.<SUP minmax_bound="true">17, 44</SUP>
Many genes involved in lipoprotein structure, secretion, and catabolism have been examined for influences on statin response. The most frequently studied of these are LDL receptor (LDLR), cholesterol ester transfer protein (CETP), and apolipoprotein E (ApoE). LDLR is a primary candidate for pharmacogenetic testing as it is directly implicated in the mechanism of statin-mediated LDL reduction. In addition, rare loss of function mutations in the gene encoding LDLR cause familial hypercholesterolemia (FH), a condition marked by strikingly elevated plasma cholesterol levels and increased risk of mortality. Statin-treated FH patients show variable responses to treatment in an LDLR mutation-dependent manner. Carriers of loss of function mutations appear more responsive than null function carriers.<SUP minmax_bound="true">45, 46, 47, 48, 49</SUP> More common, less dysfunctional LDLR genetic variants could alter response in non-FH hypercholesterolemic individuals. Several small studies suggest that this is true, although none of these associations were replicated in larger populations such as PRINCE or ACCESS.<SUP minmax_bound="true">13, 17, 50, 51</SUP>
Genes involved in hepatic regulation of LDLR expression are also hypothesized to contribute to genetic variation in statin response. LDLR expression is highly regulated by many intracellular pathways, most significantly through transcriptional regulation by the sterol regulatory element binding protein (SREBP) family of transcription factors. These are sequestered in the endoplasmic reticulum under normal conditions and, when signaled by sterol depletion, are escorted by SREBP cleavage activating protein (SCAP) into the Golgi for proteolytic activation. Variations in SREBP1, SREBP2, and SCAP have been examined for genetic associations to statin response. SREBP1 promoter variant G-36D predicted the magnitude of statin-induced ApoAI elevation in one of two studies, whereas a polymorphism in SCAP has been reported to influence response of LDLC, TC, and TG.<SUP minmax_bound="true">52, 53</SUP> Association of SCAP variation with response of multiple lipid parameters suggests validity, but the SCAP variant associations have not replicated.<SUP minmax_bound="true">53, 54</SUP> Recently, a novel protease, PCSK9, has been implicated in LDLR protein degradation and variation in this gene was associated with basal plasma LDLC.<SUP minmax_bound="true">55, 56, 57</SUP> PCSK9 gain-of-function mutations causing increased degradation of LDLR and reduced clearance of plasma LDL were associated with higher pretreatment LDLC levels and attenuated statin-mediated reduction of LDLC.<SUP minmax_bound="true">55, 56</SUP> This newly discovered element of the lipoprotein regulatory system may be an important contributor to statin response.
The most closely examined gene connected with lipoprotein structure and metabolism is that encoding ApoE, a protein that resides on triglyceride-rich atherogenic lipoproteins as well as HDL, and interacts with LDLR to promote hepatic particle clearance. Three major variants of this form have been identified;
3, which encodes wild type protein, 2 and 4, both of which contain one non-synonymous variant causing altered protein function. The presence of variant isoforms causes altered particle clearance and is associated with hyperlipidemic states. Many studies have examined the contributions of these isoforms to statin-mediated response. The consensus is that 4 carriers appear to have attenuated lipid-lowering response, and 2 carriers have enhanced response.<SUP minmax_bound="true">13, 17, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70</SUP> Few studies have examined the effects of variations in genes encoding other apolipoproteins. There are reports that polymorphisms in APOB, the gene encoding the major structural protein in both LDL and triglyceride-rich lipoproteins, may alter LDLC response; whereas variation in the APOAI gene, encoding the major structural protein in HDL, may alter HDLC response.<SUP minmax_bound="true">13, 17, 50, 62, 66, 71, 72, 73</SUP>Modulation of lipoprotein lipid content is carried out by a series of proteins that act through mechanisms of inter-lipoprotein lipid transfers and lipolysis. CETP catalyzes the transfer of lipids between HDL and triglyceride-rich lipoproteins. Variation in this gene, particularly a TaqIB variant located within intron 1, has been reported to be associated with plasma HDL concentration and CVD risk under basal conditions.<SUP minmax_bound="true">74, 75</SUP> Many studies have examined the influence of the TaqIB variant on statin response. Initial reports suggesting that TaqIB is associated with variation in HDLC response and CVD disease end points have failed to replicate.<SUP minmax_bound="true">13, 76, 77, 78</SUP> Boekholdt et al.<SUP minmax_bound="true">75</SUP> recently performed a meta-analysis, using data from 13 500 subjects across 10 trials but could not confirm this result. Although this variant may have a minor effect on response to statin therapy, it appears that the magnitude of the contribution is too small to easily distinguish from other sources of variation. Other lipid-modulating genes that have been studied include microsomal triglyceride transfer protein (MTP), which also encodes for a lipid transfer protein, and the lipase genes, lipoprotein lipase (LPL) and hepatic lipase (LIPC). Single studies have reported that a promoter variant in MTP is associated with triglyceride response and recurring risk of coronary events, an LPL variant may alter TC or HDLC response, and an LIPC variant may alter HDLC response.<SUP minmax_bound="true">79, 80, 81, 82, 83, 84, 85</SUP>
Removal of cholesterol from peripheral cells requires active transport and is mediated by ABCA1. Preliminary pharmacogenetic analysis suggests that an ABCA1 haplotype may attenuate ApoAI response to statin treatment.<SUP minmax_bound="true">86</SUP>
ABCG5 and ABCG8 are transporters that act within the enterocyte to limit intestinal absorption, and within the hepatocyte to increase biliary excretion.<SUP minmax_bound="true">87, 88</SUP> Rare, loss of function mutations in these genes lead to sitosterolemia, a disorder characterized by hypercholesterolemia caused by increased intestinal cholesterol absorption and decreased biliary secretion.<SUP minmax_bound="true">89</SUP> Intestinal absorption of cholesterol is tightly regulated in conjunction with hepatic cholesterol synthesis to maintain homeostatic levels of systemic cholesterol.<SUP minmax_bound="true">90</SUP> LDLC resistance to statin therapy may be related to a compensatory increase in cholesterol absorption via upregulation of ABCG5 and ABCG8.<SUP minmax_bound="true">91, 92, 93</SUP> As such, these are major candidates for pharmacogenetic examination. To date, there is suggestive evidence that an ABCG8 variant is associated with LDLC response to statin treatment.<SUP minmax_bound="true">13, 17, 92</SUP>
Pathways governing the pleiotropic or non-lipid response to statin therapy are under active investigation. There is evidence that these pathways contribute to statin-induced reductions in coronary events and mortality but their importance in this regard is not well established. As these pathways are incompletely characterized, it is difficult to identify candidate genes that may have pharmacogenetic influence. Nevertheless, a handful of studies have examined variation in putative response genes. As the mechanisms involved are independent of lipid-lowering effects, response in these studies tends to be defined by number of post-treatment coronary events or degree of atherosclerotic lesion progression. Replicate studies have been reported for only three genes: ACE, IGTB3, and TLR4. The gene encoding angiotensin converting enzyme (ACE) plays a role in regulation of blood pressure. An insertion/deletion variant within the ACE gene has been reported to predict risk of recurring coronary events following statin treatment in several studies.<SUP minmax_bound="true">94, 95, 96</SUP> The IGTB3 gene encodes the platelet-specific fibrinogen receptor and variation in this gene has been associated under basal conditions with elevated risk of coronary disease.<SUP minmax_bound="true">94, 97</SUP> This increase in risk was reported to be abolished by statin treatment.<SUP minmax_bound="true">94, 97</SUP> The TLR4 gene encodes for toll-like receptor 4, which is involved in innate immunity. Carriers of a non-synonymous TLR4 variant (D299G) demonstrated reduced risk of post-treatment coronary events.<SUP minmax_bound="true">98, 99</SUP> None of the mechanisms for these associations is known. In short, studies regarding the role of variation in genes involved in non-lipid-lowering response to statins have not been definitive. A clearer understanding of the components of these pathways is required before a suitable list of candidate genes can be examined.
Although variation in genes involved in cholesterol biosynthesis or lipoprotein production and clearance are established contributors to variation in basal concentrations of plasma lipids, studies to date have not definitely demonstrated a link to variation in statin-mediated lipid response. The lack of success of studies aimed at discovering the genetic variation contributing to therapeutic response is due partly to methodological limitations, most notably lack of sufficient statistical power, and partly to the nature of the goal. Statin-mediated lipid reduction is a complicated system with many components. Individual examination of isolated genetic variants is unlikely to provide useful answers. Recent studies have moved to a multiple candidate gene strategy, but these have not yet examined the influence of interactions between these variations. This approach, as well as genome-wide association or linkage studies using larger populations, may prove more productive than previous studies.
Top of page Genetic influences on statin-mediated adverse drug reactions
A particularly promising application of pharmacogenetics in statin therapy is the use of genetic pre-screening to identify subjects susceptible to adverse drug reactions. Although statin therapy is relatively well tolerated, with mild dyspepsia and gastrointestinal discomfort accounting for the majority of reported adverse reactions, more severe reactions can occur, notably severe hepatic dysfunction (
0.7% frequency) and skeletal myopathy (0.1% frequency). These have been reported with all statins to varying degrees.<SUP minmax_bound="true">11</SUP> Although these events are rare, given the exceptional number of patients prescribed these drugs, the occurrences in real terms are considerable and serious enough to cause public concern.<SUP minmax_bound="true">100</SUP> Muscle biopsies from patients indicated that myopathy is defined on the cellular level by mitochondrial dysfunction, increased lipid storage, and ragged red muscle fibers.<SUP minmax_bound="true">101</SUP> Approximately 2?5% of myopathies lead to rhabdomyolosis, characterized by muscle destruction and myoglobinuria.<SUP minmax_bound="true">101</SUP> As myopathy is linked to elevated creatine kinase levels, monitoring of creatine kinase concentrations during statin treatment is recommended to identify cases before symptoms become severe.<SUP minmax_bound="true">102</SUP> Notably, not all cases of statin-induced myopathy are accompanied by elevated creatine kinase.<SUP minmax_bound="true">103</SUP>Risk and severity of myopathy with stain treatment, unlike statin efficacy, is related to magnitude and duration of systemic exposure.<SUP minmax_bound="true">102</SUP> Thus, the risk for this complication is more susceptible to any source of variation in pharmacokinetic parameters. Systemic exposure is affected by many non-genetic factors, including age, renal and hepatic function, smoking status, and concomitant medications. It is also likely to be affected by polymorphisms in the genes involved in pharmacokinetic handling of statins. Many of the studies discussed above examined the effect of variation in these genes on plasma AUC. Although this is not an appropriate predictor of efficacy, it may be a useful predictor of toxicity. Additionally, some of the transporters noted above, SLC22A6/8 and ABCB1, in particular, may be involved in transport of statins into skeletal muscle cells.<SUP minmax_bound="true">14, 17, 104</SUP> Current reports suggest that the SLC16 family of transporters, responsible for lactic acid transport across plasma membrane into muscle, also interacts with statins.<SUP minmax_bound="true">105, 106</SUP> Inhibition of SLC16A4, which is primarily expressed in skeletal but not cardiac myocytes, paralleling the site of clinical myopathy, is reported to attenuate statin-induced myopathy in vitro.<SUP minmax_bound="true">106, 107</SUP> This may be a good candidate gene for statin toxicity studies.
Pharmacogenetic analysis of statin myopathy is limited partly because the physiological mechanisms underlying statin-induced myopathy are not well understood. Myopathy may be linked to mitochondrial dysfunction, particularly within Complex I of the respiratory chain.<SUP minmax_bound="true">108</SUP> Initially thought to result from intracellular cholesterol depletion and alterations of membrane fluidity, myopathy is currently thought to be independent of cholesterol depletion and has been suggested to depend on depletion of Coenzyme Q, also a downstream derivative of mevalonate.<SUP minmax_bound="true">109</SUP> Coenzyme Q10 is a member of the mitochondrial respiratory chain as well as having involvement in maintaining cellular health.<SUP minmax_bound="true">110, 111</SUP> Both plasma and muscular concentrations of Coenzyme Q10 are depleted by statin therapy.<SUP minmax_bound="true">112, 113</SUP> Although Coenzyme Q10 supplements during statin therapy have been advocated for relieving muscle discomfort and reducing risk for myopathy, there is no conclusive evidence for their effectiveness.<SUP minmax_bound="true">114</SUP> Alternatively, statins may cause myopathy by impairing intracellular ion pumps, altering Na+ and Ca2+ concentrations, and impairing cell conductance or inducing apoptosis.<SUP minmax_bound="true">115, 116</SUP>
Pharmacogenetic analysis of statin-induced myopathy is also limited by availability of cases. With a frequency of 0.1%, even the largest of the statin trials include only a handful of serious myopathy cases, making genetic association studies very difficult to power. Analyses to date are based on case studies or examination of a small number of subjects. Variation in CYP2D6 has been associated with decreased adherence to simvastatin therapy, although the metabolic significance of this enzyme in simvastatin clearance is under debate.<SUP minmax_bound="true">23, 25</SUP> The presence of the
4 allele of ApoE is associated with decreased adherence.<SUP minmax_bound="true">67</SUP> A case of rhabdomyolysis-induced death from cerivastatin exposure may have been caused by a loss of function mutation in CYP2C8 resulting in increased systemic exposure.<SUP minmax_bound="true">117</SUP> The SLCO1B1<SUP minmax_bound="true">*</SUP>15 haplotype, causing decreased hepatic uptake of pravastatin, is over-represented in subjects presenting with rhabdomyolysis.<SUP minmax_bound="true">118</SUP>Novel methods of recruitment will be necessary to compile populations better served to address the question of genetic variation in statin-induced ADR. Recently Wilke et al.<SUP minmax_bound="true">119</SUP> examined 68 cases of statin-induced myopathy within a retrospective case?control study of patients identified from the Marshfield Clinic patient population for association between CYP3A genotype and risk or severity of myopathy, as measured by serum CK level. Results indicated that severity but not risk was associated with the presence of two copies of the CYP3A5<SUP minmax_bound="true">*</SUP>3 haplotype. Ruano et al. examined 19 SNPs from 10 candidate genes involved in vascular homeostasis for association with statin-induced elevated serum CK levels and reported that polymorphisms in the gene for angiotensin II receptor 1(AGTR1), a multifunctional protein, and the gene encoding nitric oxide synthase 3 (NOS3), an enzyme involved in smooth muscle contraction and platelet aggregation, were associated with serum CK.<SUP minmax_bound="true">120</SUP> These two studies represent the first major examinations of genetic variation in statin-induced myopathy.
Top of page Summary of clinical implications
Although, statin pharmacogenetics is still in its infancy, the consensus based on current evidence is that genetic testing to assist in selecting a particular statin and/or dosing regimen is not clinically warranted. There are several reasons for this opinion. Statin therapy is so well tolerated in the majority of patients that the need to use additional testing to avoid deleterious side effects is relatively small. Moreover patients should be monitored for signs of hepatic or muscle toxicity to avoid more severe adverse reactions. In addition, patients on statin therapy should have lipid levels monitored so that submaximal response can be improved by increased dosage. Another major concern is that nearly all studies of pharmacogenetics of statin response reported to date have been underpowered and replication of positive results is lacking. Finally, most of these studies have examined associations of individual genetic polymorphisms with treatment response. Although genetic differences may prove a significant source of response variability, this is unlikely to be due to any single gene. Rather, the compound effects of multiple genetic variants are more likely to be responsible. Most studies have examined a small number of variants in a single or small number of genes ? often one variant in one gene ? and have reported the resultant outcomes based on small patient populations.
There are, however, other considerations that support the potential value of statin pharmacogenetic information. Identification of genes and genetic variants that influence statin responsiveness holds promise for identifying molecular components of physiological lipid and inflammatory pathways that mediate statin effects. However, the most important clinical benefits of statin pharmacogenetic knowledge would be based on identifying a set of genotypes that aid in predicting the outcomes of statin treatment in terms of reduced risk for cardiovascular events, coupled with reduced risk for serious ADRs. As studies to date have documented at most a 30% statin-mediated reduction in risk for CVD events and stroke, it would be valuable to have information on potential genetic influences on the likelihood of a beneficial outcome. However, a meaningful analysis of the genetic determinants responsible for statin efficacy and toxicity will require more comprehensive analysis of genome-wide variations in multiple large-scale populations. Until such information is available for either efficacy or safety, there is little reason to consider the use of genetic testing for guiding statin treatment.
Top of page Notes
Duality of interest
None declared.