Pharmacogenomics of Rosuvastatin: A Glocal (Global+Local) African Perspective and Expert Review on a Statin Drug

Abstract

The incidence of cardiovascular diseases (CVDs) in African populations residing in the African continent is on the rise fueled by both a steady increase in CVD risk factors and comorbidities such as human immunodeficiency virus/acquired immune deficiency syndrome (HIV/AIDS), tuberculosis, and parasitic diseases such as bilharzia. Statins are recommended together with lifestyle changes in the treatment of hypercholesterolemia and overall reduction of cardiovascular events. Rosuvastatin in particular is an attractive candidate in the management of CVDs in African populations often plagued with multimorbidities owing to both its potency and low drug-to-drug interaction potential. In this expert review, we describe the pharmacogenetics of rosuvastatin and how it may instrumentally affect the African populations. We describe polymorphisms in the candidate genes, ABCG2, SLCO1B1, CYP2C9, APOE, PCSK9, LDLR, LPA, and HMGCR, and their role in the potency and safety of rosuvastatin therapy. We report on qualitative and quantitative differences in the distribution of genetic variants that affect efficacy and toxicity of rosuvastatin. These differences are observed across world populations (Caucasian, European, and Asian) as well as within African populations. Finally, we advocate for extensive pharmacogenetic studies in African populations that take into account the genetic diversity of intra-African ethnic groups and the genetic differences between African populations and other global populations, with a collaborative and collective aim to provide effective and safe use of rosuvastatin in management of CVD in Africa. Our key thesis presented in this innovation field analysis is that rosuvastatin precision medicine can serve as a veritable Glocal (Global and Local) model to offer pharmacogenetic-guided optimal therapeutics for the public in both developing and developed regions of the world.

Introduction

Rosuvastatin is one of the six statins available on the market and is indicated in the treatment of hypercholesterolemia, hypertriglyceridemia, hyperlipidemia, mixed dyslipidemia, and homozygous familial hypercholesterolemia (Astra Zeneca, 2015). Rosuvastatin was first approved by the U.S. Food and Drug Administration (FDA) in 2003 for use in the treatment of hypercholesterolemia (Astra Zeneca, 2015; FDA, 2015). Similar to other statins, rosuvastatin acts by inhibiting the production of mevalonic acid, a rate-limiting step in the biosynthesis of cholesterol. X-ray crystallography studies show that the statins' characteristic pharmacophores, which resemble 3-hydroxy-3-methylglutaryl coenzyme A (HMGCoA), inhibit the HMGCoA reductase (HMGCR) by attaching to the enzyme's active site and sterically inhibiting the binding of its substrate HMGCoA and hence the production of mevalonate (Istvan et al., 2001).

The overall effect is a reduction in intracellular cholesterol levels in hepatocytes, which upregulates the synthesis of low-density lipoprotein receptors (LDLRs) and their expression on hepatocyte membranes, consequently resulting in a drop in circulating total cholesterol and total LDL cholesterol (LDL-C) levels in blood. Numerous studies have reported on the potency of statins in the reduction of total cholesterol in patients with hypercholesterolemia (Chasman et al., 2012; Ford et al., 2002; McTaggart et al., 2001; Olsson, 2001; Paoletti et al., 2001; SEARCH Collaborative Group et al., 2008); however, statins also present with other pleiotropic effects beyond cholesterol reduction, making them the most prescribed cholesterol-reducing class of drugs globally (Ioannidis, 2014).

Some of the pleiotropic effects include reduction of triglycerides (Grosser et al., 2004; Paoletti et al., 2001), statin-induced upregulation of endothelial nitric oxide (Lefer, 2002; Mayer et al., 2007), and reduced leukocyte–endothelial cell interactions (Pruefer et al., 1999). Mayer et al. (2007) showed that rosuvastatin in particular has anti-inflammatory, antithrombotic, and antioxidative properties. Based on these findings, rosuvastatin is now indicated for patients with primary hypercholesterolemia, mixed dyslipidemia, hypertriglyceridemia, homozygous familial hypercholesterolemia, slowing progression of atherosclerosis, pediatric patients aged 10–17 years with heterozygous familial hypercholesterolemia, and also for risk reduction of myocardial infarction, stroke, and arterial revascularization procedures in patients without evident coronary heart disease, but displaying multiple risk factors (Astra Zeneca, 2015; FDA, 2015).

Following these studies confirming the potency of statins against hypercholesterolemia, the World Health Organization (WHO, 2007) recommended the use of satins in primary prevention of cardiovascular events in patients with established risk factors such as total cholesterol levels above 8 mmol/L (320 mg/dL), persistent raised blood pressure, type 1 and type 2 diabetes, and renal failure or impairment. WHO also recommends the use of statins in the treatment of individuals with established coronary heart disease and cerebrovascular disease. Consequently, these recommendations have been adopted across the world (Klug, 2012; Reiner et al., 2011; Stone et al., 2014).

Given that cardiovascular diseases (CVDs) are the world's number one cause of death (WHO, 2016), causing 17.5 million deaths in 2012 alone, statins have become the most prescribed drug across the globe. In the United States alone, an estimated 33 million people without established CVDs have an increased risk for heart attack or stroke in the next 10 years (Go et al., 2014), an estimated 25% of American adults aged ≥45 years are therefore expected to be on statins if WHO recommendations are followed; thus, based on these estimations, statin prescriptions are projected to reach the trillion mark globally by 2020 (Ioannidis, 2014).

Data on CVDs and statin use in sub-Saharan Africa are fairly limited. However, CVD is on a steady increase in this WHO region pushed by two main factors, namely an epidemiological transition and other disease conditions that increase CVD risk, such as human immunodeficiency virus/acquired immune deficiency syndrome (HIV/AIDS), tuberculosis (TB), and schistosomiasis (Dau and Holodiny, 2008; Huaman et al., 2015; Moolani et al., 2012). This epidemiological transition has its root causes primarily in increased urbanization and the rapid globalization, which has seen marked changes in lifestyles among African populations. Improved socioeconomic status also means longer life spans and hence more people are presenting with heart disease as they age.

Using data derived from large clinical and administrative databases in health maintenance settings, researchers at Kaiser Permanente (Carr et al., 2013) have consistently demonstrated that patients with HIV have a higher chance of hospitalization for CVDs, especially acute myocardial infarction, than the negative controls. HIV, which is a huge burden in sub-Saharan Africa, introduces its own host of cardiac complications such as increased incidences of inflammatory circulatory disorders, including macrovascular arteries, pulmonary hypertension, cardiomyopathy, and TB pericarditis (Dau and Holodiny, 2008; Ntsekhe and Hakim, 2005). Although antiretroviral therapy (ART) prolongs the life of HIV/AIDS patients, this is coming at a cost of increased incidence of atherosclerosis as individuals grow older. Furthermore, ART has been associated with increased incidence of dyslipidemia, insulin resistance, inflammation, changes in body fat distribution, and drug toxicities, all which are CVD risk factors (Dau and Holodiny, 2008).

In a bid to respond to the burden of CVDs in Africa, the Heart of Soweto Study was designed (Sliwa et al., 2008) to monitor, describe, and respond to the evolving burden of heart disease in a large predominantly black urban population. It was reported that heart failure was the most common primary diagnosis with black Africans more likely to be diagnosed of heart failure than other populations. This study suggested that the burden of CVD was further broadened by the presence of cardiac complications relating to TB and HIV/AIDS (e.g., TB pericardial effusion, HIV-cardiomyopathy, and dyslipidemia related to ART). Thus, the high burden of increasing CVD risk factors coupled with HIV/AIDS, TB, and schistosomiasis in the sub-Saharan African group will probably increase the need for statins to prevent and manage CVDs.

Given the large number of comorbidities and multimorbidities that plague sub-Saharan African populations, it is important that the statin of choice to prevent and manage CVDs must be potent at low concentrations and have fewer drug-to-drug complications. Chapman and McTaggart (2002) describe the ideal statin as one that has potent, yet reversible, inhibition of the target HMGCR both in vivo and in vitro and should have selective uptake by hepatocytes accompanied by minimal tissue accumulation. The statin should also have minimal drug–drug interactions. Rosuvastatin, a third-generation statin, shows the highest potency as measured by both lipid-lowering effects, inhibition of HMGCR (Istvan et al., 2001; McTaggart, 2003; White, 2002) and high selectivity for hepatocytes, compared with the other members of the statin family (Brown et al., 2001; Buckett et al., 2000; Nezasa et al., 2002, 2003).

The most attractive aspect of the pharmacology of rosuvastatin for the sub-Saharan African region is its low lipophilicity, which lowers the need for extensive metabolism by CYP450 isozymes, reducing its dependence on CYP enzymes for elimination. In vitro investigations show no significant rosuvastatin inhibition of CYP1A2, CYP2C9, CYP2C19, CYP2D6, CYP2E1, and CYP3A4 in human hepatic microsomes (McCormick et al., 2000). Additionally, in vivo studies using ketoconazole, fluconazole, and erythromycin further point to the absence of metabolism of rosuvastatin by CYP450 isozymes, with only a minimal 10% of orally administered rosuvastatin under the metabolism of CYP2C9 to N-desmethylrosuvastatin (Cooper et al., 2002, 2003; McTaggart et al., 2001). Thus, rosuvastatin is unlikely to be associated with clinically significant metabolic drug-to-drug interactions, although caution has been issued when coadministered with protease inhibitors such as ritonavir and lopinavir and the immunosuppressant, cyclosporine (Astra Zeneca, 2015; FDA, 2015).

The low possibility of drug–drug interactions makes rosuvastatin an attractive candidate for prevention and management of CVDs in sub-Saharan African populations as these populations are often burdened with multiple conditions and infections concomitantly, which would require use of different drugs. In one of the few studies on comorbidity in African populations, nearly 50% of the study participants recruited in South Africa had a comorbidity, while 14% of the 18,856 consultations had multimorbidity (two or more conditions) (Lalkhen and Mash, 2015), strengthening the option of rosuvastatin in the management and prevention of CVDs in the African continent.

Set on this background, this article sets out to give an overview on the pharmacokinetics (PKs) and pharmacogenomics of rosuvastatin with special emphasis on African populations. The pharmacogenomics of statins has been comprehensively reviewed elsewhere (Alfonsi et al., 2016; Leusink et al., 2016). This review aims to describe the pharmacogenomics of statin-induced myopathy (SIM) and differences in rosuvastatin exposure among different ethnic populations and how these differences play an important role in the use of rosuvastatin in the prevention and management of CVDs in sub-Saharan African populations.

Literature Review

Literature for this review was accessed from PubMed/Medline and Google Scholar databases. The main keywords used to search for literature included rosuvastatin used in combination with the following words; pharmacology, pharmacokinetics, African, black, genetics, Pharmacogenomics, pharmacogenomics, and HIV/AIDS. Other keywords used to search the databases include CVD in Africa, pharmacogenomics, statins, SLCO1B1, ABCG2, PCSK9, HMGCR, LDLR, LPA, and ApoE. The review was limited to articles that could be accessed in full and only articles written in English were considered.

Pharmacogenomic Considerations

Pharmacogenomics is an arm of genetics that deals with how genetic variants affect the response to therapeutic agents (Alfonsi et al., 2016; Dandara et al., 2014; Leusink et al., 2016; Masimirembwa and Hasler, 2013; Verschuren et al., 2012; Voora and Ginsburg, 2012). The main areas of pharmacogenomics that aid in our understanding of how individuals respond to drugs encompass PKs (which is how the individual disposes of the drug), pharmacodynamics (which is the effect of the drug on the individual) and lastly the underlying disease mechanism (Alfonsi et al., 2016; Voora and Ginsburg, 2012), and pharmacoepigenetics. The very aim of pharmacogenomic studies at both individual and population levels is to understand the contribution of the underlying genetics in a patient on how they respond to therapeutic drug treatment. Understanding pharmacogenomics is likely to provide for ways to consider appropriate doses of therapeutic drugs for patients with different genotypes. This review will focus on issues pertaining to the safety and potency of rosuvastatin use in African populations.

The Genetics of Rosuvastatin and Drug-by-Gene Interactions

Since their entry into the market in 1987, statins have outperformed all other lipid-lowering drugs and now rank as the most prescribed lipid-lowering drugs globally (Gelissen and Brown, 2012; Marrero et al., 2012). They owe their huge success to both their potency and relative safety. Rosuvastatin has a favorable safety profile, which is comparable with the other statins (Luvai et al., 2012). The rare adverse drug reactions associated with rosuvastatin therapy include nosebleeds, cold-like symptoms, headaches, insomnia, and digestive system problems (Astra Zeneca, 2015). There have been reports on increased incidence of hypoglycemia, liver dysfunction, and myopathy associated with rosuvastatin therapy (GISSI-HF Investigators, 2008; Kjekshus et al., 2007; Ridker et al., 2010).

However, the most reported side effect of rosuvastatin therapy is by far SIM (Schacter, 2004; Shannon et al., 2012, 2013; Shepherd et al., 2004), which is reported in nearly 1% of the elderly (above 65 years) (Chasman et al., 2012). Incidence of SIM in monitored clinical trials, as measured by a creatine kinase level 10 times the upper normal limit, ranges from about 0.01 to 0.3% with rhabdomyolysis incidence at ∼0.003–0.01% (Armitage, 2007; Backes et al., 2008; Chasman et al., 2012; Law and Rudnicka, 2006; SEARCH Collaborative Group et al., 2008). Although the Justification for the Use of Statins in Primary Prevention (JUPITER) trail reported no significant difference in myopathy incidence between the placebo and rosuvastatin groups (Chasman et al., 2012), adverse reports from the U.S. FDA indicate that the incidence of SIM in clinical practice is far higher and in some cases as high as 9–20% in selected outpatient cases (Bellosta et al., 2004; Fernandez et al., 2011). These discrepancies are fueled primarily by the controlled nature of recruitment of patients in clinical trials where patients with high susceptibility to SIM are usually excluded.

Second, the lack of a reliable diagnostic definition of SIM also may lower classification of reported cases. Traditionally, a creatine kinase level 10 times the upper normal limit has been used in trials as a biochemical marker of SIM, but reports of myopathy have been made in patients without elevated creatine kinase (Niemi et al., 2011), putting its reliability as a biomarker in question.

SIM has been shown to affect statin adherence especially among the elderly (Rossi and McLeod, 2009) and is thought to account for up to 40% of patients lost to follow-up while on statin therapy (Shannon et al., 2012). Although the exact mechanism of SIM remains largely vague, studies have consistently shown that statin dose and incidence of SIM are linearly related as incidence of SIM is linked to systemic drug exposure (Lee et al., 2005; Pasanen and Backman, 2006). Therefore, factors that affect the disposition of rosuvastatin are likely to also affect the incidence of SIM in patients. These factors include gender, age, body–mass index, and renal and hepatic function, as well as genetics. The scope of this review is limited to the genetic factors that affect safety and potency of rosuvastatin therapy in African populations.

Rosuvastatin primarily acts in the liver where it lowers the synthesis of cholesterol by inhibiting the rate-limiting step. The disposition of rosuvastatin is through bile excretion in hepatic cells, thus hepatocytes are involved both in its disposition and potency. Hepatocytes therefore present an epicenter of most of the PKs and pharmacodynamics of rosuvastatin. A lot of information generated around the potency and safety of rosuvastatin centers around the enzymes involved in its metabolism and transporters involved in its uptake and excretion from hepatocytes (Table 1). Thus, genetic variation in the following genes, SLCO1B1, ABCG2, CYP2C9, HMGCR, LPA, LDLR, PCSK9, and ApoE, is important in how individuals are likely to respond to rosuvastatin treatment (Table 1).

Table 1.

Global Distribution of Genetic Variants That Affect Rosuvastatin Efficacy and Safety (1000 Genomes Project Consortium, 2015)

			African populations			European populations		Asian populations
Gene	Polymorphism	Allele	Yoruba (Nigeria)	Luhya (Kenya)	African-American	Finnish	British	Chinese	Japanese
SLCO1B1	c.521T>C p:V174A (rs4149056)	C	0.07	0.02	0.01	0.18	0.14	0.14	0.12
	c.388 A>G p:N130D (rs2306283)	G	0.25	0.16	0.19	0.44	0.36	0.22	0.34
	g-11187G>A (rs4149015)	A	0.01	0.03	0.01	0.10	0.02	0.12	0.12
	g.89595 T>C (rs12317268)	C	0.23	0.23	0.14	0.28	0.14	0.46	0.35
	g.7344 A>G (rs2900478)	A	0.07	0.01	0.01	0.19	0.15	0.13	0.11
ABCG2	c.421C>A;p:Q141K (rs2231142)	A	0.07	0.00	0.00	0.09	0.14	0.31	0.32
CYP2C9	^*2:c.430C>T p:C144R (rs1799853)	T	0.00	0.00	0.04	0.08	0.09	0.00	0.00
	^*3 c.1075A>C p:I359 L (rs1057910)	C	0.00	0.00	0.02	0.06	0.07	0.04	0.02
HMGCR	9.331648A>T (rs17244841)	T	0.09	0.09	0.10	0.01	0.03	0.00	0.02
	g.27506T>G (rs17238540)	G	0.10	0.09	0.10	0.01	0.02	0.00	0.00
	c.^*372T>C (rs12916)	C	0.41	0.19	0.32	0.45	0.43	0.52	0.53
	g.23092 A>G (rs3846662)	A	0.13	0.03	0.04	0.55	0.55	0.49	0.47
	c.1866T>C (rs1047433)	C	0.43	0.46	0.38	0.58	0.67	0.34	0.22
LDLR	g.7250 G>T (rs651720)	T	0.12	0.18	0.13	0.06	0.07	0.00	0.00
	g.11084147 G>A (rs1672123)	A	0.01	0.00	0.00	0.39	0.27	0.07	0.06
ApoE	c.526c>T; R176C (rs7412)	T	0.13	0.05	0.11	0.07	0.08	0.11	0.05
LPA	g.160589086A>G (rs10455872)	G	0.02	0.00	0.00	0.04	0.08	0.00	0.00
PCSK9	c.137G>T; R46 L (rs11591147)	G	1.0	1.0	1.0	0.04	0.02	0.00	0.00

Genes Involved in the Disposition of Rosuvastatin

Solute-like carrier 1B1

Solute-like carrier 1B1 (SLCO1B1) is a gene that encodes organic anion transporter polypeptide 1B1 (OATP1B1) and is highly expressed in the liver and intestines (Niemi et al., 2011). Rosuvastatin has minimal passive diffusion across the cell membrane due to its low lipophilicity and is transported across the cells primarily by OATP1B1 and to a lesser extent OATP1B3 and OATP2B1.SLCO1B1 (Ho et al., 2006; Kitamura et al., 2008; Kopplow et al., 2005; Lee et al., 2005; Pasanen and Backman, 2006).

The SLCO1B1 gene has generated a lot of interest in the pharmacogenomics of statins and in particular rosuvastatin. Polymorphisms in SLCO1B1 have been shown to affect transporter activity of OATP1B1 by either increasing or decreasing its activity. Due to the overwhelming evidence of the involvement of SLCO1B1 in statin potency and safety among Europeans, the Clinical Pharmacogenomics Implementation Consortium (Ramsey et al., 2014) issued a guideline for SLCO1B1 and simvastatin-induced myopathy (Ramsey et al., 2014). These guidelines are based on continual public literature for SLCO1B1 polymorphisms and their effects on transport activity SLCO1B1 single-nucleotide polymorphisms (SNPs).

SLCO1B1 c.521T>C (rs4149056, SLCO1B1*5, p.V174A) has consistently shown association with SIM (Birmingham et al., 2015; Chasman et al., 2012; DeGorter et al., 2013; Pasanen and Backman, 2006; Pasanen et al., 2006; Voora and Ginsburg, 2012). The SLCO1B1 521C allele results in decreased expression of OATP1B1 (Tirona et al., 2001) and hence decreased transporter activity, which ultimately increases systemic exposure to rosuvastatin increasing the risk of SIM. SLCO1B1 c.521C allele has frequencies of 15–20%, 15%, and 2% among Caucasians, Asians, and sub-Saharan Africans, respectively (Niemi et al., 2011).

However, genetic data emerging from sub-Saharan Africa show that the frequency of the SLCO1B1c.521C polymorphisms is variable, confirming the heterogeneity of African populations, ranging from as low as 0% to as high as 21% (Aklillu et al., 2011; Ngaimisi et al., 2013; 1000 Genomes Project Consortium, 2015; Soko et al., 2016), yet even in populations with 0%, PK parameters of rosuvastatin are equally variable (Soko et al., 2016). Thus, huge variability in the distribution of SLCO1B1 c.521T>C among African populations cautions the use of this variant to predict rosuvastatin exposure for all African populations. There is need to evaluate among African populations whether groups with high incidence of SLCO1B1 c.521T>C have this as a marker of rosuvastatin exposure and need to investigate for novel variants explaining rosuvastatin variability in population groups that do not have this variant.

The SLCO1B1 c.521C allele is found with the *1B G allele (rs2306283, c.388A>G, p.N130D) in the SLCO1B1*15 haplotype, which has been consistently associated with reduced transporter activity and hence increased SIM risk (Pasanen et al., 2006). However, on its own, SLCO1B1*c.388A>G has yielded conflicting results with some reports of increased activity and others of decreased activity. In the SEARCH study (SEARCH Collaborative Group, 2008), c.388A>G had borderline significance (p = 0.03) between the G allele and reduced risk of SIM on simvastatin patients. This SNP occurs at a frequency of 22% in Caucasian populations and an average of 78% in sub-Saharan African populations (1000 Genomes Project Consortium, 2015; Ramsey et al., 2014); whether these differences would result in lowered incidence of SIM in African populations compared with European populations remains to be seen.

Another SLCO1B1 haplotype is *17, which consists of SLCO1B1 c.521C, c.388G, and the G allele of rs4149015 (g.-11187G>A). This haplotype has also been associated with reduced transporter activity (Ramsey et al., 2014). The haplotype frequency as reported by the CPIC is 2% among sub-Saharan Africans, 0% among African-Americans, and 21% in Caucasians. Once again, the reduced frequency of this haplotype and the SLCO1B1*5 and *15 haplotypes in African populations gives the impression that there is reduced risk of SIM among all sub-Saharan African populations.

Two intronic SNPs, SLCO1B1 g.89595T>C (rs4363657) and g.73414A>G (rs12317268), were identified from the JUPITER study/trial as having significance in attenuating absolute LDL-C-lowering response (Chasman et al., 2012). SLCO1B1 rs4363657 is in strong linkage disequilibrium with SLCO1B1c.521T>C (r² = 0.97) and has been reported above with genome-wide significance in a genome-wide association study (GWAS) with SIM patients (SEARCH Collaborative Group, 2008) with an odds ratio of 4.5 per copy of the C allele, 16.9 for the CC homozygous when compared with the TT wild type. There are significant interethnic differences in the frequency of SLCO1B1 g.89595T>C polymorphism among sub-Saharan African populations (50%) and European populations (90%) (1000 Genomes Project Consortium, 2015).

The second intronic SNP, SLCO1B1 g.73414A>G, occurs at comparable frequencies between African (20%) and European populations (MAF 18%) (1000 Genomes Project Consortium, 2015). Other SLCO1B1 variants reported to influence statin response in general include SLCO1B1 g.89670T>A (rs2900478), an intronic SNP reported in linkage distribution with SLCO1B1 c.521T>C associated with attenuated lowering of absolute LDL-C (Postmus et al., 2012). The minor g.89670A allele occurs at frequencies of 17%, 1% among Caucasians and Africans, respectively (1000 Genomes Project Consortium, 2015). SLCO1B1 c.463C>A (*4 rs11045819, p.P115T) has also been associated with increased transporter activity when the c.463A allele is found together with the c.388G allele SLCO1B1 *14 haplotype (Ramsey et al., 2014). This increased transporter activity is associated with enhanced LDL-C-lowering effect. SLCO1B1*14 has a frequency of 86% in Europeans and 94% in Africans (1000 Genomes project Consortium, 2015).

The differences in SLCO1B1 polymorphisms associated with attenuated LDL-C lowering of rosuvastatin between Caucasians and Africans give the impression that the incidence of SIM among African populations should be relatively low and also the LDL-C-lowering effect of rosuvastatin should be higher than in Caucasian populations. However, this may not be so as many other genes are involved in the overall potency and safety profile of rosuvastatin. Second, there is a possibility of the role of other variants that have larger frequencies in African populations and/or even novel variants in these candidate genes that are specific to African populations and are yet to be identified. African populations have diverse frequencies for SLCO1B1 c.521C allele, a trend that may possibly play out with all the other genes involved in rosuvastatin disposition and potency, thus affecting how the various African ethnic groups would respond to rosuvastatin therapy.

There is need for more resources to genetically characterize the largely heterogeneous African populations if precision medicine is to be achieved. Furthermore, research needs to move from extrapolating African data from African-Americans or from one African ethnic group for other groupings. African populations need to be treated as separate entities, not lumped together as one big cohort because they are a heterogeneous group.

ATP Binding Cassette G2

ATP binding cassette G2 (ABCG2) encodes the efflux transporter breast cancer resistance (BCRP) transporter. BCRP is constitutively expressed in the enterocytes and on the canalicular membrane of hepatocytes (Mo and Zhang, 2012). In the intestinal lumen, BCRP actively pumps rosuvastatin back into the lumen from the enterocytes, hence limiting absorption of the drug from the gastrointestinal tract. In the hepatocytes, BCRP pumps rosuvastatin from hepatocytes out for biliary excretion. Candidate gene approaches have consistently associated polymorphism ABCG2 rs2231142 (c.421C>A, p. Gln141Lys) with increased risk of SIM (Birmingham et al., 2015; DeGorter et al., 2013; Keskitalo et al., 2009; Lee et al., 2005; Zhang et al., 2006). The ABCG2 c.421A variant occurs in the intracellular domain of the polypeptide (Ni et al., 2010) and results in reduced protein expression (Kobayash et al., 2005; Kondo et al., 2004) and hence reduced efflux capability in enterocytes and hepatocytes among individuals carrying the 421A allele, consequently systemic exposure, as marked by area under the curve (AUC) of plasma drug concentration–time curve, of rosuvastatin increases (Birmingham et al., 2015; DeGorter et al., 2013; Keskitalo et al., 2009; Lee et al., 2005). The ABCG2 c.421A allele occurs at a frequency of 10–15% among Caucasians, 25–35% among Asians, and 0–5% among sub-Saharan Africans (Mao and Unadkat, 2005; Niemi et al., 2011).

Following these reports, the FDA Public Health Advisory for Crestor issued a revision for the dosage of rosuvastatin in patients with Asian ancestry; the starting dose was revised downward from 20 mg once daily to 5 mg once daily in patients of Asian ancestry (Astra Zeneca, 2015; FDA, 2005). A more detailed investigation of the frequency of ABCG2 c.421C>A among different African ethnicities conducted in our research group shows that there are no significant differences in the frequency of the A allele between nine ethnic groups, with an average MAF of 2%. Thus, among sub-Saharan populations, ABCG2 rs2231142 is likely to play a minor role in the pharmacogenomics of rosuvastatin in these populations.

In a GWAS (Chasman et al., 2012), ABCG2 g.11244T>C (rs2199936), which is in linkage disequilibrium with ABCG2 c.421C>A (rs2231142) (r² = 0.81) and occurred at a minor allele frequency of 11% in this study population, was the only ABCG2 variant found to be associated with attenuated absolute LDL-C lowering at genome-wide significance of p < 5 × 10⁻⁸ (Chasman et al., 2012). This variant was the most significant locus associated with absolute LDL-C reduction in this study, with the mean per allele effect being −5.2 mg/dL corresponding to observed median LDL-C reductions of −53 mg/dL (rs2199936GG), −59 mg/dL (rs2199936AG), and −5.2 mg/dL (rs2199936AA) (Chasman et al., 2012). This variant occurs at a similar MAF of 91% in both European and African populations (1000 Genomes Project Consortium, 2015).

In the same study, g.11393T>C (rs1481012), which is in linkage disequilibrium with ABCG2 rs2199936 and occurred at an MAF of 11% in this study population, had a significant association with fractional LDL-C reduction (p < 5.8 × 10⁻¹⁹) and baseline LDL-C reduction (p < 1.6 × 10⁻⁴⁷) (Chasman et al., 2012). Fractional LDL-C was calculated as absolute difference in cholesterol divided by baseline volume. This variant is not common in African populations occurring at 1% (1000 Genomes Project Consortium, 2015), while the MAF in Europeans is 9%.

CYP2C9

Due to its relative hydrophilicity, rosuvastatin requires minimal metabolism from the cytochrome P450 (CYP450) isozymes before biliary elimination (McTaggart, 2003); however, in vitro studies in human microsomes have shown that CYP2C9 is the major metabolizing isozyme, converting rosuvastatin into the less potent metabolite, N-desmethylrosuvastatin, with minor contribution of CYP2C19 (Cooper et al., 2003; McCormick et al., 2000). Cooper et al. (2002) showed that the antifungal potent inhibitor of CYP2C9, fluconazole, resulted in at least 14% increase in AUC of rosuvastatin. Thus, CYP2C9*2 and CYP2C9*3, which occur in frequencies of 7–9% (Masimirembwa and Hasler, 2013) in African populations and produce an inactive enzyme (Aithal et al., 1999; Furuya et al., 1995), are likely to affect bioavailability of rosuvastatin. In 2009, Galleli published a case report describing the development of rhabdomyolysis in a female Caucasian patient in response to rosuvastatin, which attributed the adverse drug reaction to CYP2C9 saturation due to the concomitant intake with warfarin. The patient was moved to fluvastatin, which is not metabolized by CYP2C9, and the rhabdomyolysis cleared (Galleli et al., 2009).

Genes Involved in Rosuvastatin Potency

HMGCoA reductase

HMGCR is the primary target protein in statin therapy, thus an important candidate gene in any statin therapy. However, to date, no polymorphism in this important pharmacogene has been able to reach genome-wide significance in rosuvastatin response (Chasman et al., 2012). HMGCR g.331648A>T (rs17244841) was first associated with attenuated statin response in a study of 1536 participants taken from the Pravastatin Information/CRP Evaluation (PRINCE) study (Chasman et al., 2004), with an associated 22% reduction in total cholesterol lowering after pravastatin treatment in heterozygous patients compared with A/A homozygotes. rs1724481A>T is an intronic SNV, the minor T allele has a frequency of 0.02 in Europeans (1000 Genomes Project Consortium, 2015) and is more frequent in individuals of African ancestry (0.09). Chasman also identified another SNV on exon 15, HMGCR g.27506T>G (rs17238540), which was in close linkage disequilibrium (r² > 0.9) with rs17244841. The minor G allele frequency was 0.035, heterozygotes had a 22.3% reduction in total cholesterol lowering and 19.0% lower reduction in LDL-C after taking 40 mg of pravastatin daily.

These results were repeated in the cholesterol and pharmacogenetics (CAP) study (Simon et al., 2006), which had 326 African-American participants and 596 Caucasian participants (Krauss et al., 2008). A haplotype H7 was defined by Chasman as a combination of the minor alleles of rs17244841 and rs17238540 and HMGCR g.331648T and HMGCR g.27506G, respectively (Krauss et al., 2008). Another HMGCR g.23092A>G (rs3846662) SNP was reported, which is involved in the regulation of alternative splicing of HMGCR at exon 13 (Medina, 2010), resulting in an HMGCR protein devoid of exon 13 transcription. This alternative splicing at exon 13 is used as a marker of statin therapy as its expression is correlated with the magnitude of plasma cholesterol, apolipoprotein B, and triglyceride reductions in statin therapy (Medina, 2010).

In the CAP study, H7 carriers had a 20.5% smaller reduction of total cholesterol and a 24.4% smaller reduction of LDL-C in response to simvastatin when compared with noncarriers (Chasman et al., 2004). In the CAP study, however, H7 frequency was 0.03 in the Caucasians and 0.06 in the African-Americans. Interestingly, these attenuated responses were only observed in the African-American population. Another haplotype H2, which is defined by Krauss as a combination of common alleles across 11 tag SNPs within the HMGCR, occurred at a frequency of 0.32 in African-Americans and 0.02 in Caucasians in the CAP study (Krauss et al., 2008). African-Americans in the CAP study who were carriers of H2 had 10.3% smaller reductions of total cholesterol and 10% smaller reductions in LDL-C (Krauss e al., 2008). Notably, H2/H7 carriers had even further reductions in LDL-C response than individuals who carried either H2 or H7 haplotypes.

Although the CAP study was done on patients on simvastatin, it highlights major differences in response to statin therapy based on differences in genetic factors among different ethnic groups. H7 and H2 all occur in higher frequencies in African-American groups and likewise reduce the potency of simvastatin; for the African population, it becomes necessary to evaluate the potential reduction in potency these haplotypes may have on rosuvastatin therapy, this may affect dose regimens for individuals of African ancestry as starting doses derived from studies in predominantly Caucasian participants may potentially have lower efficacy in some African populations.

Low-Density Lipoprotein Receptor

LDLR is a surface glycoprotein, which binds to proteins apoB and apoE in lipoproteins. The binding of these lipoproteins and the subsequent internalization of LDL-C suppress cholesterol synthesis in the cell through suppression of HMGCR. The use of statins increases the expression of LDLR, which in turn increases the clearance of LDL-C from the plasma. In candidate gene studies (Shepherd et al., 1999), two closely linked SNVs (r² = 0.997) showed that the minor allele C44857T had a significantly, yet modest, larger LDL-C-lowering response (p = 0.03). Thus, the haplotype, C44857T [T]-A44964G [A], had a lower risk of the primary endpoints and a lower risk of CVD events.

In the CAP study (Mangravite et al., 2010), significant allelic differences were observed between black and white populations. The CAP study patients were randomized into simvastatin and placebo groups. Using 5 tag SNVs (rs14148, rs1433099, rs7254521, rs5742911, and rs2738467), Mangravite et al. (2010) showed that as separate SNVs, none of these tag SNVs was associated with statin-mediated response, but the haplotype L5, which contained only major alleles for all tag SNVs, was associated with attenuated statin-mediated changes in total cholesterol, apoB, and non-high-density lipoprotein (HDL) cholesterol levels. Interestingly, this observation was only seen in black patients where L5 had a frequency of 21% compared with 3% in whites.

Individuals with both LDLR L5 and the HMGCR H2/H7 haplotypes showed an even more attenuated response to simvastatin. The combined contribution of these haplotypes was 3–6% (Mangravite et al., 2010) in lipid-lowering effect of simvastatin, whereas these combined haplotypes contributed less than 1% in white patients owing to their rarity. These differences at candidate gene level strengthen the need to investigate dose requirements in African populations as the presence of these haplotypes can lead to lower therapeutic effect and consequently the need for higher dosages among African populations. As these haplotypes have shown consistent effect in two different statins, it is possible to infer that should African populations have a higher prevalence of LDLR L5 and or HMGCR H2/H7 than the Caucasian populations on which rosuvastatin dose was determined, it is possible that doses may need to be adjusted for rosuvastatin therapy in African patients with mixed dyslipidemia or hypercholesterolemia.

To date, no SNV in LDLR has reached genome-wide significance in GWASs on patients on rosuvastatin. The most significant SNV in the LDLR locus was g.7250G>T (rs6511720) in the JUPITER study (Chasman et al., 2012). This intronic variant, which had a minor allele frequency of 9% in this predominantly Caucasian population, had a significant association with percentage reduction in LDL-C (−2.6% per allele; p = 0.005) (Chasman et al., 2012). The SNV g.7250G>T has a higher frequency in individuals of African ancestry (Table 1) and as such can be expected to cause an increased attenuated response in these populations.

Apolipoprotein E

The apolipoprotein E (ApoE) gene encodes a major apolipoprotein of chylomicrons, ApoE, which is a serum glycoprotein found predominantly in circulating chylomicrons and chylomicron remnants such as very LDL (VLDL), LDL, and HDL. Apo E is essential for the normal catabolism of triglyceride-rich constituents with its most significant role being acting as a ligand for receptor-mediated clearance of chylomicron and VLDL remnants. The amino terminal of the protein is responsible for binding of ApoE to the LDLR, and the carboxyl terminal mediates the binding of Apo E to the surface of lipoproteins. The Apo E protein is highly polymorphic with three proteins, namely E2, E3, and E4. These three proteins are determined by three haplotypes ɛ2, ɛ3, and ɛ4 (Utterman et al., 1977). Three homozygous genotypes arise (E2/E2, E3/E3, and E4/E4) and three heterozygote genotypes also occur (E2/E3, E3/E4, and E4/E4). These haplotypes arise from three main variants, rs7412T-rs429358T (E2), rs7412C-rs429358T (E3), and rs7412C-rs429358C (E4) (Fullerton et al., 2000). The affinity for binding to LDLR and lipoprotein particles differs among the three proteins (Anoop et al., 2010) with the E4 protein having the highest affinity for VLDL, hence reducing its circulation in blood faster and consequently increasing circulating cholesterol levels.

Candidate gene studies in other statins such as atorvastatin, simvastatin, and pravastatin show an attenuated response to LDL-C lowering for individuals with the E3 allele compared with those with the E2 allele (Chasman et al., 2012; Chu et al., 2015; Yang et al., 2005). These studies together with GWAS show that E2 (rs7412) genotype had the greatest response to statin effect, whereas the E4 (rs429358) had the lowest response (Chasman et al., 2012). In the JUPITER GWAS, which is specific to rosuvastatin, the ApoE locus was one of the four loci that had significant association with LDL-C reduction at a genome-wide significance of p < 5 × 10⁻⁸ (Chasman et al., 2012). ApoE c.526C>T p:Arg176Cys (rs7412), which occurred at an MAF of 15% in this predominantly European population, was significant at both fractional LDL-C percentage reduction (p < 5.8 × 10⁻¹⁹) and baseline LDL-C reduction (p < 1.6 × 10⁻⁴⁷).

Elimination of ApoE c.526C>T from the analysis reduced genome-wide significance of these loci to levels below p < 5 × 10⁻⁸, implying this locus has a significant effect on the potency of rosuvastatin as measured by LDL-C-lowering effect. The frequency of the minor T allele of rs7412 is 8% globally, 10% in African populations, and 6% in European populations (1000 Genomes Project Consortium, 2015), while that of c.388T>C p:Cys130Arg (rs429358) C allele is 15% globally, 27% for Africans, and 16% for Europeans (1000 Genomes Project Consortium, 2015).

E3 is the most commonly expressed protein globally, but interethnic differences in frequencies of ApoE variants have been reported. In Africa, the frequency of the three ApoE haplotypes is ɛ2 (12%), ɛ3 (71%), and ɛ4 (18%) (Atadzhanov et al., 2013). However, due to the heterogeneous nature of African genetics, these figures vary considerably across the continent, with ɛ3 being almost exclusively expressed among a West African group, the Wolof (93.9%) (Zekraoui et al., 1997); ɛ2 has significantly high frequencies in the Mossi (37.5%) of Burkina Faso, Ewe (31.6%) of Togo, and the Malinke (23.3%) of Togo (Zekraoui et al., 1997). High frequencies of ɛ4 were also observed in the Fon (29.4%) of Benin, Tutsi (38%) of Rwanda, and some pygmies (66%) of Central African Republic (Zekraoui et al., 1997). This mixed expression of different ApoE haplotypes among different African ethnicities supports the need for a more comprehensive investigation of the pharmacogenomics of rosuvastatin among intra-African ethnicities.

In the JUPITER study, (Chasman et al., 2012) reported that another ApoE locus, the intronic variant, c.-245T>C (rs71352238), with an MAF of 10% in this population, showed borderline association (p < 2.9 × 10⁻⁴) with LDL-C lowering. This variant independently showed significant association with lowering of absolute LDL-C. g.-245T>C occurs at a frequency of 13% in Europeans and 1% in Africans (1000 Genomes Project Consortium, 2015). However, although c.-245T>C occurs at a frequency of 0.06 in African-Americans and 0.01 in African-Caribbeans in Barbados (1000 Genomes Project Consortium, 2015), it is rare in African populations residing in the African continent, with a minor allele frequency of 0.00 in the Yoruba from Nigeria, a 0.00 frequency in the Luo of Kenya, and 0.00 in the Mandinka of The Gambia (1000 Genomes Project Consortium, 2015). These MAF differences among global African populations further confirm the need to steer away from extrapolating genetic data of Africans in Africa from African-American populations.

Lipoprotein A

The gene lipoprotein A (LPA) encodes the expression of the serine proteinase lipoprotein (a), commonly called LP (a). LP (a) consists of three polypeptides, one molecule of apolipoprotein apoB100, an LDL, and a high-molecular-weight glycoprotein (a). LP (a) undergoes proteolytic cleavage to yield lipoprotein fragments that promote thrombogenesis by attaching to atherosclerotic lesions and is therefore a strong independent risk factor of atherosclerosis (Khera et al., 2014; Kronenberg, 2014). In a study in 1394 Brazilian patients with clinical diagnosis of angina pectoris and stable angina (Santos et al., 2014), the LPA intronic variant g.161010118A>G (rs10455872) G allele was associated with higher incidence of coronary lesions. LPA g.161010118A>G rs10455872 had an MAF of about 5% in the JUPITER study and it was one of the four loci that had a signal above the genome-wide significance of p < 5 × 10⁻⁸.

The study confirmed attenuated response to a statin as had been shown earlier by Postmus et al. (2014) in patients with g.161010118A>G. The mean allele change in LDL-C in rosuvastatin therapy in the JUPITER study was 6.2 mg/dL, corresponding to observed median LDL-C reductions of −55, −48, and −46 mg/dL with increase in minor allele copies (Chasman et al., 2012). Although the study did not measure LP (a) levels, it could be inferred that LP (a) concentrations could be highly correlated with lower rosuvastatin effects on plasma LDL-C. The LPA rs g.161010118A>G has an MAF of ∼7% in Europeans and 0% in African populations (1000 Genomes Project Consortium, 2015). There is very little knowledge on the actual frequency of this variant among the genetically diverse African populations as its impact on rosuvastatin and in fact on any other statin's potency in these populations remains largely unexplored.

Proprotein Convertase Subtilisin–Kexin 9

Proprotein convertase subtilisin–kexin 9 (PCSK9) is a protease highly expressed in the liver. PCSK9 is involved in the post-transcriptional regulation of hepatic LDLR expressed on cell surfaces; it binds to LDLR (Kwon et al., 2008; Schmidt et al., 2008; Zhao et al., 2006), modifying the receptor for lysosomal degradation. As a result, gain-of-function mutations increase the degradation of LDLR and hence increase plasma cholesterol levels, resulting in hypercholesterolemia (Abifadel et al., 2003; Maxwell and Breslow, 2004; Sun et al., 2005), whereas loss-of-function mutations decrease degradation of LDLR and the overall effect is hypocholesterolemia and a resultant protective effect against CVD (Benn et al., 2010; Cohen et al., 2006; Langstead et al., 2016). Individuals with loss of function of PCSK9 are expected to have a greater response to statins as LDLR degradation is reduced.

GWASs on patients taking statins have shown that PCSK9 c.137 G>T, p: Arg46Leu, R46 L (rs11591147) has a significant effect on baseline LDL-C concentrations and hence the overall effect of the statin. In a GWAS of patients from the Prospective Study of Pravastatin in the Elderly at Risk (PROSPER) trial, PCSK9 R46 L had a 10% lower LDL-C level (p = 3.62 × 10⁻¹²) and a 9% reduced risk of vascular disease (Postmus et al., 2014). Similarly, PCSK9 R46 L reached a genome-wide significance of p = 4.7 × 10⁻⁹ for baseline LDL-C concentrations in the JUPITER study (Chasman et al., 2012). PCSK9 is a loss-of-function mutation that confers protection against hypercholesterolemia in individuals with the T allele. The T allele occurs at a frequency of 2% in Europeans and is rare among individuals of African ancestry (Table 1).

In a study involving 669 Americans (Feng et al., 2016), PCSK9 occurred at a frequency of 0.002 in African-Americans. As expected, carriers of PCSK9 rs11591147 T allele had a greater reduction in LDL-C levels in response to statins, 55.6% (p = 0.0064) than individuals with the G allele. They also identified population-specific genetic variants; rs11583680 occurred at a frequency of 2% in Americans of African descent and was absent in European Americans. Another variant, c.2037 C>A, p:C697X (rs28362261), had a frequency of 2% in African-Americans, but was absent in Americans of European descent. The loss-of-function variant, rs28362261, was associated with statin response among these African-Americans (p = 0.0064). This polymorphism is a C to A transversion at nucleotide position 2037 that results in truncation of the protein by 14 amino acids (Cohen et al., 2005). It has a frequency of 1.4% in Nigerians from a Yoruba-speaking rural community (Cohen et al., 2006) and a frequency of 3.7% in women from Zimbabwe of African descent (Hooper et al., 2007).

Loss-of-function mutations in PCSK9 seem to have the most significant effect on statin therapy in patients, hence loss-of-function mutations may also affect rosuvastatin therapy in patients with hypercholesterolemia in Africa. However, the loss-of-function mutation, R46 L, is rare in African individuals, yet the C679X mutation is more common, occurring at a rate of 1–3.7% (Cohen et al., 2006; Feng et al., 2016; Hooper et al., 2007) in patients of African descent. It can be inferred that C697X may have a larger effect on rosuvastatin response in African populations than does R46 L, which is found in Caucasian and not African populations.

Future Perspectives and Conclusion

Pharmacogenomics of rosuvastatin has been extensively studied in Caucasian populations. The drug was developed and approved on the backbone of results from clinical trials and both PK and pharmacogenomic studies in predominantly Caucasian populations. In these populations, rosuvastatin is fairly tolerated with rare adverse drug events such as SIM. The candidate genes involved in rosuvastatin SIM have also been identified using both candidate gene approaches and GWASs. Two polymorphisms are consistently associated with SIM, SLCO1B1 c.521T>C and ABCG2 c.421C>A, SLCO1B1 c.521T>C increasing SIM in all populations and ABCG2 c.421C>A in particular increasing incidence of SIM in patients of Asian individuals. However, both SLCO1B1 c.521T>C and ABCG2 c.421C>A together with CYP2C9*2 and CYP2C9*3 are rare in African populations, giving the impression that the incidence of SIM in African populations should be low.

A PK trial was conducted in 30 healthy male Zimbabweans of African descent (Soko et al., 2016); all 30 volunteers lacked both the SLCO1B1 c.521C and ABCG2 c.421A alleles associated with increased rosuvastatin plasma concentrations; however, interindividual variability was noted with a 10-fold difference in maximum plasma concentrations of rosuvastatin between the volunteer with the lowest Cmax (2.8 ng/mL) and the one with the highest Cmax (31 ng/mL). The major limitation of this study was the sample size and sample recruitment, which targeted only a small subset of African ethnicities. Variation of PKs is expected to differ among the differing ethnicities owing to differences in frequencies of polymorphisms associated with pharmacogenomics of rosuvastatin.

This study triggers the need to carry out larger and longer PK trials in African populations so as to establish the actual incidence of SIM among different African ethnic populations. The absence of these two polymorphisms in these populations could point to other polymorphisms at play among the individuals that may be unique or even novel to African populations. There is need therefore for African researchers to employ third-generation sequencing technologies to further understand the causes of these interindividual differences.

In terms of efficacy, to the best of our knowledge, there are no trials evaluating the efficacy of rosuvastatin in African populations from sub-Saharan Africa. Both candidate gene studies and GWASs consistently identify polymorphisms in the genes ApoE, LDLR, LPA, HMGCR, and PCSK9 as having an effect on the overall LDL-C-lowering ability of rosuvastatin. Again, there are significant differences in both the nature and frequency of polymorphisms in these genes between the Caucasian populations and the African populations. In studies done using the CAP study, a simvastatin trial (Krauss et al., 2008; Mangravite et al., 2010), significant differences were observed between frequencies of the haplotypes, HMGCR H2/H7 and LDLR L5, which were more prevalent in African populations and were associated with attenuated response to simvastatin. These findings confirm the need to carry out clinical trials of similar magnitude in African populations. African populations can no longer afford to just blindly adopt pharmacogenetic results from other populations if the safe and efficacious use of rosuvastatin is to be achieved.

Conclusions and Executive Take-Home Points

Overall, pharmacogenetic research in African populations on the African continent is limited. Constraints in medical record management in health facilities affect the access of data by African researchers and coupled with limited human and financial capacity, the pharmacogenetic output from the African continent remains far below that from other populations. Yet, African populations are the most genetically diverse group and a one-size-fits-all approach does not necessarily fit in all circumstances. For instance, SLCO1B1 c.521C allele has a higher frequency in East African populations (16%) (Aklillu et al., 2011; Ngaimisi et al., 2013), versus a low frequency in Southern Africa (0–5%) (Soko et al., 2016) and therefore although the SLCO1B1 c.521C polymorphism may play an important role in the pharmacogenomics of East African populations, it may play an almost negligent role in Southern African populations.

Large population studies across the continent that investigate the PKs of rosuvastatin across different African ethnicities will allow for a more comprehensive outlook on rosuvastatin therapy. Different African ethnicities also need to evaluate the frequencies of various polymorphisms involved in the pharmacogenomics of rosuvastatin to allow ease of access of data for researchers, clinicians, and policy makers alike. Deep sequencing and resequencing of candidate genes, whole exomes, and whole genomes are also necessary as novel common and rare polymorphisms may play a role in rosuvastatin therapy in African populations. This step would aid the already existing efforts in developing more efficient genotype arrays and diverse whole genome panels for accurate imputation of common variation in African populations (Gurdasani et al., 2015).

Pharmacogenomics of statins has been a productive research area over the years, but a recent systematic review (Leusink et al., 2016) showed that replication of results has been fairly low. Study of rosuvastatin PKs and pharmacogenomics in African populations offers a chance to replicate and confirm the results of previous studies and add on new findings to the current body of knowledge. Although pharmacogenomics remains important in the rosuvastatin therapy, other omics such as metabolomics, proteomics, and the microbiome need to be explored if we are to understand and improve safe and efficacious use of rosuvastatin in the treatment of dyslipidemia and hypercholesterolemia in the African continent.

Last, the recently formed African Pharmacogenomics Consortium aims to increase pharmacogenomic research in Africa and this could assist in genetic characterization of African populations. However, African governments and researchers need to define the pharmacogenomic research policy and programs for the continent by taking into account both the genetic diversity of its people and their diverse health-related issues and aim to improve treatment. Our key thesis presented in this innovation field analysis is that rosuvastatin precision medicine can serve as a veritable model to offer pharmacogenomic-guided optimal therapeutics for the public in both developing and developed regions of the world.

Footnotes

Author Disclosure Statement

The authors declare that no conflicting financial interests exist.

Abbreviations Used

References

Abifadel

, Varret

, Rabès

, et al. (2003). Mutations in pcsk9 cause autosomal dominant hypercholesterolemia. Nat Genet, 34, 154–156.

Aklillu

, Mugusi

, Ngaimisi

, et al. (2011). Frequency of the SLCO1B1 388A>G and the 521T>C polymorphism in Tanzania genotyped by a new LightCycler®-based method. Eur J Pharmacol, 67, 1139–1145.

Aithal

, Day

, Kesteven

PJL

, and Daly

. (1999). Association of polymorphisms in the cytochrome P450 CYP2C9 with warfarin dose requirement and risk of bleeding complications. Lancet, 353, 717–719.

Alfonsi

, Hegele

, and Gryn

. (2016). Pharmacogenetics of lipid-lowering agents: precision or indecision medicine?. Curr Atheroscler Rep, 18, 24.

Anoop

, Misra

, Meena

, and Luthra

. (2010). Apolipoprotein E polymorphism in cerebrovascular & coronary heart diseases. Indian J Med Res, 132, 363–378.

Armitage

. (2007). The safety of statins in clinical practice. Lancet, 370, 1781–1790.

Astra Zeneca. (2015). Crestor (rosuvastatin calcium) tablets. Prescribing Information. 1–43.

Atadzhanov

, Mwaba

, Mukomena

, et al. (2013). Frequency of ApoE, MTHFR and ACE polymorphisms in the Zambian population. BMC Res Notes, 7, 194.

Backes

, Venero

, Gibson

, et al. (2008). Effectiveness and tolerability of every-other-day rosuvastatin dosing in patients with prior statin intolerance. Ann Pharmacother, 42, 341–346.

10.

Bellosta

, Paoletti

, and Corsini

. (2004). Safety of statins: focus on clinical pharmacokinetics and drug interactions. Circulation, 109, 1150–1157.

11.

Benn

, Nordestgaard

, Grande

, Schnohr

, and Tybjarg-Hansen

. (2010). PCSK9 R46 L, low density lipoprotein cholesterol levels, and risk of ischemic heart disease: 3 independent studies and meta-analyses. J Am Coll Cardiol, 55, 2833–2842.

12.

Birmingham

, Bujac

, Elsby

, et al. (2015). Impact of ABCG2 and SLCO1B1 polymorphisms on pharmacokinetics of rosuvastatin, atorvastatin and simvastatin acid in Caucasian and Asian subjects: a class effect?. Eur J Pharmacol, 71, 341–355.

13.

Brown

CDA

, Windass

, Bleasby

, and Lauffart

. (2001). Rosuvastatin is a high affinity substrate of hepatic organic anion transporter OATPC. Atheroscler Suppl, 2, 90.

14.

Buckett

, Ballard

, Davidson

, Dunkley

, et al. (2000). Selectivity of ZD4522 for inhibition of cholesterol synthesis in hepatic versus non-hepatic cells. Atherosclerosis, 151, 41.

15.

Carr

, Meara

, Jorgensen

, et al. (2013). SLCO1B1 genetic variant associated with statin induced myopathy: a proof-of-concept study using the clinical practice research datalink. Clin Pharmacol Ther, 9, 695–701.

16.

Chapman

, and McTaggart

. (2002). Optimizing the pharmacology of statins: characteristics of rosuvastatin. Atheroscler Suppl, 2, 33–37.

17.

Chasman

, Giulianini

, MacFadyen

, Barratt

, Nyberg

, and Ridker

. (2012). Genetic determinants of statin-induced low-density lipoprotein cholesterol reduction: the justification for the use of statins in prevention: an intervention trial evaluating rosuvastatin (JUPITER) trial. Circ Cardiovasc Genet, 5, 257–264.

18.

Chasman

, Posada

, Subrahmanyan

, Cook

, Stanton

, and Ridker

. (2004). Pharmacogenetic study of statin therapy and cholesterol reduction. JAMA, 291, 2821–2827.

19.

Chu

, Guilianini

, Barratt

, Nyberg

, Chasman

, and Ridker

. (2015). Pharmacogenetic determinants of statin-induced reductions in C-reactive protein. Circ Cardiovasc Genet, 5, 58–65.

20.

Cohen

, Pertsemlidis

, Kotowski

, Graham

, Garcia

, and Hobbs

. (2005). Low LDL cholesterol in individuals of African descent resulting from frequent nonsense mutations in PCSK9. Nat Genet, 37, 161–165.

21.

Cohen

, Boerwinkle

, Mosley

Jr. , and Hobbs

. (2006). Sequence variations in PCSK9, low LDL, and protection against coronary heart disease. N Engl J Med, 354, 1264–1272.

22.

Cooper

, Martin

, Dane

, Warwick

, Raza

, and Schneck

. (2003). Lack of effect of ketoconazole on the pharmacokinetics of rosuvastatin in healthy subjects. Br J Clin Pharmacol, 55, 94–99.

23.

Cooper

, Martin

, Dane

, Warwick

, Schneck

, and Cantarini

. (2002). The effect of fluconazole on the pharmacokinetics of rosuvastatin. Eur J Pharmacol, 58, 527–531.

24.

Dandara

, Swart

, Mpeta

, Wonkam

, and Masimirembwa

. (2014). Cytochrome P450 pharmacogenetics in African populations: implications for public health. Expert Opin Drug Metab Toxicol, 10, 769–785.

25.

Dau

and Holodiny

. (2008). The relationship between HIV infection and CVD. Curr Cardiol Rev, 4, 203–218.

26.

DeGorter

, Tirona

, Schwarz

, et al. (2013). Clinical and pharmacogenetic predictors of circulating atorvastatin and rosuvastatin concentrations in routine clinical care. Circ Cardiovasc Genet, 6, 400–408.

27.

Feng

, Wei

, Chung

, et al. (2016). The effect of genetic variation in PCSK9 on the LDL-cholesterol response to statin therapy. Pharmacogenomics J [Epub ahead of print]; DOI: 10.1038/tpj.2016.3.

28.

Fernandez

, Spatz

, Jablecki

, and Phillips

. (2011). Statin myopathy: a common dilemma not reflected in clinical trials. Cleve Clin J Med, 78, 393–403.

29.

Food and Drug Administration (FDA). (2015). Highlights of prescribing information. www.accessdata.fda.gov/drugsatfda_docs/label/2010/021366s016lbl.pdf Last access: July 31, 2016.

30.

Food and Drug Administration. (2005). FDA Alert on Crestor (rosuvastatin calcium), 3/2005.FDA ALERT [03/2005].

31.

Ford

, Blauw

, Murphy

, et al. (2002). A prospective study of pravastatin in the elderly at risk (PROSPER): screening experience and baseline characteristics. Curr Control Trial Cardiovasc Med, 3, 8.

32.

Fullerton

, Clark

, Weiss

, Nickerson

, Taylor

, and Stengard

. (2000). Apolipoprotein E variation at the sequence haplotype level: implications for the origin and maintenance of a major human polymorphism. Am J Hum Genet, 67, 881–900.

33.

Furuya

, Fernandez-Salguero

, Gregory

, et al. (1995). Genetic polymorphism of CYP2C9 and its effect on warfarin maintenance dose requirement in patients undergoing anticoagulation therapy. Pharmacogenetics, 5, 389–392.

34.

Galleli

, Ferraro

, Spagnuolo

, Rende

, Mauro

, and De Sarro

. (2009). Rosuvastatin-induced rhabdomyolysis probably via Cyp2C9 saturation. Drug Metabol Drug Interact, 24, 83.

35.

Gelissen

, and Brown

. (2012). Predicting response to statins by pharmacogenetic testing. Pharmacogenomics, 13, 1223–1225.

36.

GISSI-HF Investigators. (2008). Effect of rosuvastatin in patients with chronic heart failure (the GISSI-HF trial): a randomised, double-blind, placebo-controlled trial. Lancet, 372, 1231–1239.

37.

, Mozaffarin

, Roger

, et al. (2014). Heart disease and stroke statistcs-2014 update. A report from the American Heart Association. Circulation, 129, e29–e290.

38.

Grosser

, Erdmann

, Hmmerle

, et al. (2004). Rosuvastatin upregulates the antioxidant defense protein heme oxygenase-1. Biochem Biophys Res Commun, 325, 871–876.

39.

Gurdasani

, Carstensen

, Tekola-Ayele

, et al. (2015). The African Genome Variation Project shapes medical genetics in Africa. Nature, 517, 327–332.

40.

, Tirona

, Leake

, et al. (2006). Drug and bile acid transporters in rosuvastatin hepatic uptake: function, expression, and pharmacogenetics. Gastroenterology, 130, 1793–1806.

41.

Hooper

, Marais

, Tanyanyiwa

, and Burnett

. (2007). The C679X mutation in PCSK9 is present and lowers blood cholesterol in a Southern African population. Atheroscleorsis, 193, 445–448.

42.

Huaman

, Hensen

, Ticona

, Sterling

, and Garvy

. (2015). Tuberculosis and cardiovascular disease: linking the epidemics. Trop Dis Travel Med, 1, 10.

43.

Genomes Project Consortium. (2015). A global reference for human genetic variation. Nature, 526, 68–74.

44.

Ioannidis

. (2014). More than a billion people taking statins?: potential implications of the new cardiovascular guidelines. JAMA, 311, 463–464.

45.

Istvan

, Deisenhofer

, Eisenberg

, et al. (2001). Structural mechanism for statin inhibition of HMG-CoA reductase. Science, 292, 1160–1164.

46.

Keskitalo

, Zolk

, Fromm

, Kurkinen

, Neuvonen

, and Niemi

. (2009). ABCG2 polymorphism markedly affects the pharmacokinetics of atorvastatin and rosuvastatin. Clin Pharmacol Ther, 86, 197–203.

47.

Khera

, Everett

, Caulfield

, et al. (2014). Lipoprotein(a) concentrations, Rosuvastatin therapy, and residual vascular risk: an analysis from the JUPITER Trial (Justification for the Use of Statins in Prevention: an Intervention Trial Evaluating Rosuvastatin). Circulation, 129, 635–642.

48.

Kitamura

, Maeda

, Wang

, and Sugiyama

. (2008). Involvement of multiple transporters in the hepatobiliary transport of rosuvastatin. Drug Metab Dispos, 36, 2014–2023.

49.

Kjekshus

, Apetrei

, Burrios

, et al. (2007). Rosuvastatin in older patients with systolic heart failure. N Engl J Med, 357, 2248–2261.

50.

Klug

. (2012). South African dyslipidemia guideline consensus statement. S Afr Med J, 102, 178–187.

51.

Kobayash

, Ieri

, Hirota

, et al. (2005). Functional assessmet of ABCG2 (BCRP) gene polymorphisms to protein expression in human placenta. Drug Metab Dispos, 33, 94–101.

52.

Kondo

, Suzuki

, Itoda

, et al. (2004). Functional analysis of SNPs variants of BCRP/ABCG2. Pharmacol Res, 21, 1895–1903.

53.

Kopplow

, Letschert

, Walter

, and Keppler

. (2005). Human hepatobiliary transport of organic anions analyzed by quadrupole-transfected cells. Mol Pharm, 68, 1031–1038.

54.

Krauss

, Mangravite

, Smith

, et al. (2008). Variation in the 3-hydroxyl-3-methylglutaryl-coenzyme A reductase gene is associated with racial differences in low-density lipoprotein cholesterol response to simvastatin treatment. Circulation, 117, 1537–1544.

55.

Kronenberg

. (2014). Lipoprotein (a). There's life in the old dog yet?. Circulation, 129, 619–621.

56.

Kwon

, Lagace

, McNutt

, Horton

, and Deisenhofer

. (2008). Molecular basis for LDL receptor recognition by PCSK9. Proc Natl Acad Sci U S A, 105, 1820–1825.

57.

Lalkhen

, and Mash

. (2015). Multimorbidity in non-communicable diseases in South African primary healthcare. S Afr Med J, 105, 134–138.

58.

Langstead

, Nordestgaard

, Benn

, Tybjaerg-Hansen

, and Kamstrup

. (2016). PCSK9 R46 L loss of function mutation reduces lipopreotein(a), LDL-cholesterol and risk of aortic valve stenosis. J Clin Endocrinol Metab, 24, 1–9.

59.

Law

, and Rudnicka . (2006). Statin safety: a systematic review. Am J Cardiol, 97, S52–S60.

60.

Lee

, Ryan

, Birmingham

, et al. (2005). Rosuvastatin pharmacokinetics and pharmacogenetics in white and Asian subjects residing in the same environment. Clin Pharmacol Ther, 78, 330–341.

61.

Lefer

. (2002). Statins as potent antiinflammatory drugs. Circulation, 106, 2041–2042.

62.

Leusink

, Onland-Moret

, de Bakker

, et al. (2016). Seventeen years of statin pharmacogenetics: a systematic review. Pharmacogenomics, 17, 163–180.

63.

Luvai

, Mbagaya

, Hall

, and Barth

. (2012). Rosuvastatin: a review of the pharmacology and clinical effectiveness in cardiovascular disease. Clin Med Insights Cardiol, 6, 17–33.

64.

Mangravite

, Medina

, Cui

, et al. (2010). Combined influence of LDLR and HMGCR sequence variation on lipid-lowering response to simvastatin. Arterioscler Thromb Vasc Biol, 30, 1488–1492.

65.

Mao

, and Unadkat

. (2005). Role of the breast cancer resistance protein (ABCG2) in drug transport. Cell, 7, 118–133.

66.

Marrero

, Bloom

, and Adashi

. (2012). Noncommunicable diseases. JAMA, 307, 2037–2038.

67.

Masimirembwa

, and Hasler

. (2013). Pharmacogenetics in Africa, an opportunity for appropriate drug dosage regimens: on the road to personalized healthcare. CPT Pharmacometrics Syst Pharmacol, 2, e45.

68.

Maxwell

, and Breslow

. (2004). Adenoviral-mediated expression of Pcsk9 in mice results in a low-density lipoprotein receptor knockout phenotype. Proc Natl Acad Sci U S A, 101, 7100–7710.

69.

Mayer

, Gruber

, Landi

, et al. (2007). Rosuvastatin reduces interleukin-6-induced expression of C-reactive protein in human hepatocytes in a STAT3- and C/EBP-dependent fashion. Int J Clin Pharmacol, 45, 319–327.

70.

McCormick

, McKillop

, Butters

, et al. (2000). ZD4522–an HMGCoA reductase inhibitor free of metabolically mediated drug interactions: metabolic studies in human in vitro systems. J Clin Pharmacol, 40, 1055.

71.

McTaggart

. (2003). Comparative pharmacology of rosuvastatin. Atheroscler Suppl, 4, 9–14.

72.

McTaggart

, Buckett

, Davidson

, et al. (2001). Preclinical and clinical pharmacology of rosuvastatin, a new 3-hydroxy-3-methylglutaryl coenzyme A reductase inhibitor. Am J Cardiol, 87, 28–32.

73.

Medina

. (2010). The relationship between HMGCR genetic variation, alternative splicing, and statin efficacy. Discov Med, 9, 495–499.

74.

, and Zhang

. (2012). Human ABCG2: structure, function, and its role in multidrug resistance. Int J Mol Biol, 3, 1–27.

75.

Moolani

, Bukhman

, and Hotez

. (2012). Neglected tropical diseases as hidden causes of cardiovascular diseases. PLoS Negl Trop Dis, 6, e1499.

76.

Nezasa

, Higaki

, Matsamura

, et al. (2002). Liver-specific distribution of rosuvastatin in rats: comparison with pravastatin and simvastatin. Drug Metab Dispos, 30, 1158–1163.

77.

Nezasa

, Higaki

, Takeuchi

, Nakano

, and Koike

. (2003). Uptake of rosuvastatin by isolated rat hepatocytes: comparison with pravastatin. Xenobiotica, 33, 379–388.

78.

Ngaimisi

, Habtewold

, Minzi

, et al. (2013). Importance of ethnicity, CYP2B6 and ABCB1 genotype for efavirenz pharmacokinetics and treatment outcomes: a parallel-group prospective cohort study in two sub-Saharan Africa populations. PLoS One, 8, 7.

79.

, Bikadi

, Rosenberg

, and Mao

. (2010). Structure and function of the human breast cancer resistance protein (BCRP/ABCG2). Curr Drug Metab, 11, 603–617.

80.

Niemi

, Pasanen

, and Neuvonen

. (2011). Organic anion transporting polypeptide 1B1: a genetically polymorphic transporter of major importance for hepatic drug uptake. Pharmacol Rev, 63, 157–181.

81.

Ntsekhe

, and Hakim

. (2005). Impact of human immunodeficiency virus infection on cardiovascular disease in Africa. Circulation, 112, 3602–3607.

82.

Olsson

. (2001). Statin therapy and reductions in low-density lipoprotein cholesterol: initial clinical data on the potent new statin rosuvastatin. Am J Cardiol, 87, 33–36.

83.

Paoletti

, Fahmy

, Mahla

, Mizan

, and Southworth

. (2001). Rosuvastatin demonstrates greater reduction of low-density lipoprotein cholesterol compared with pravastatin and simvastatin in hypercholesterolaemic patients: a randomized, double-blind study. J Cardiovasc risk, 8, 383–390.

84.

Pasanen

, and Backman

. (2006). Frequencies of single nucleotide polymorphisms and haplotypes of organic anion transporting polypeptide 1B1 SLCO1B1 gene in a Finnish population. Eur J Clin Pharmacol, 62, 409–415.

85.

Pasanen

, Neuvonen

, and Niemi

. (2006). SLCO1B1 polymorphism markedly affects the pharmacokinetics of simvastatin acid. Pharmacogenet Genomics, 16, 873–879.

86.

Postmus

, Trompet

, Deshmukh

, et al. (2014). Pharmacogenetic meta-analysis of genome-wide association studies of LDL cholesterol response to statins. Nat Commun, 5, 5068.

87.

Pruefer

, Scalia

, and Lefer

. (1999). Simvastatin inhibits leukocyte endothelial cell interaction and protects against inflammatory processes in normocholesterolemic rats. Arterioscler Thromb Vasc Biol, 19, 2894–2900.

88.

Ramsey

, Johnson

, Caudle

, et al. (2014). The clinical pharmacogenetics implementation consortium guideline for SLCO1B1 and simvastatin-induced myopathy: 2014 update. Clin Pharmacol Ther, 96, 1–6.

89.

Reiner

, Catapano

, De Baker

, et al. (2011). ESC/EAS Guidelines for the management of dyslipidaemias. Eur Heart J, 32, 1769–1818.

90.

Ridker

, Macfadyen

, Nordestgaard

, et al. (2010). Rosuvastatin for primary prevention among individuals with elevated high-sensitivity c-reactive protein and 5% to 10% and 10% to 20% 10-year risk. Implications of the justification for use of statins in prevention: an intervention trial evaluating rosuvastatin (JUPITER) trial for “intermediate risk.”. Circ Cardiovasc Qual Outcomes, 3, 447–452.

91.

Rossi

, and McLeod

. (2009). The pharmacogenetics of statin therapy. When the body aches, the mind will follow. J Am Coll Cardiol, 54, 1617–1618.

92.

Santos

PCJL

, Bueno

, Lemos

, Krieger

, and Pereira

. (2014). LPA rs10455872 polymorphism is associated with coronary lesions in Brazilian patients submitted to coronary angiography. Lipids Health Dis, 13, 74.

93.

Schacter

. (2004). Chemical, pharmacokinetic and pharmacodynamic properties of statins: an update. Fundam Clin Pharmacol, 19, 117–125.

94.

Schmidt

, Beyer

, Bensch

, et al. (2008). Secreted proprotein convertase subtilisin/kexin type 9 reduces both hepatic and extrahepatic low-density lipoprotein receptors in vivo. Biochem Biophys Res Commun, 370, 634–640.

95.

SEARCH Collaborative Group, Link

, Parish

, et al. (2008). SLCO1B1 variants and statin induced myopathy—a genome wide study. N Engl J Med, 359, 789–799.

96.

Shannon

, John

, Ferrara

, Best

, and Allen

. (2012). Statin induced myopathy. US Pharmacist, 37, 55–59.

97.

Shannon

, John

, Parihar

, Allen

, and Ferrara

. (2013). A clinical review of statin-associated myopathy. J Pharm Tech, 29, 219–230.

98.

Shepherd

, Blauw

, Murphy

, et al. (1999). The design of a prospective study of pravastatin in the elderly at risk (PROSPER). Am J Cardiol, 84, 1192–1197.

99.

Shepherd

, Hunninghake

, Stein

, et al. (2004). Safety of rosuvastatin. Am J Cardiol, 94, 882–888.

100.

Simon

, Lin

, Hulley

, et al. (2006). Phenotypic predictors of response to simvastatin therapy among African-Americans and Caucasians: the cholesterol and pharmacogenetics (CAP) study. Am J Cardiol, 97, 843–850.

101.

Sliwa

, Wilkinson

, Hansen

, et al. (2008). Spectrum of heart disease and risk factors in a black urban population in South Africa (the Heart of Soweto Study): a cohort study. Lancet, 371, 915–922.

102.

Soko

, Dandara

, Ramesar

, Kadzirange

, and Masimirembwa

. (2016). Pharmacokinetics of rosuvastatin in 30 healthy Zimbabwean individuals of African ancestry. Br J Clin Pharmacol, 2, 326–328.

103.

Stone

, Robinson

, Lichtenstein

, et al. (2014). 2013 ACC/AHA guideline on the treatment of blood cholesterol to reduce atherosclerotic cardiovascular risk in adults. A report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines. J Am Coll Cardiol, 63, 2889–2934.

104.

Sun

, Eden

, Tosi

, et al. (2005). Evidence for effect of mutant PCSK9 on apolipoprotein B secretion as the cause of unusually severe dominant hypercholesterolaemia. Hum Mol Genet, 14, 1161–1169.

105.

Tirona

, Leake

, Merino

, and Kim

. (2001). Polymorphisms in OATP-C identification of multiple allelic variants associated with altered transport activity among European and African Americans. J Biol Chem, 276, 35669–35675.

106.

Utterman

, Hees

, and Steinmetz

. (1977). Polymorphism of apolipoprotein E and occurrence of dysbetalipoproteinemia in man. Nature, 269, 604–607.

107.

World Health Organization (WHO). (2007). Prevention of Cardiovascular Disease. Geneva.

108.

World Health Organization (WHO). (2016). Cardiovascular Disease. Geneva.

109.

White

. (2002). A review of the pharmacologic and pharmacokinetic aspects of rosuvastatin. J Clin Pharmacol, 42, 963–970.

110.

Verschuren

JJW

, Trompet

, Wessels

JAM

, et al. (2012). A systematic review on pharmacogenetics in cardiovascular disease: is it ready for clinical application?. Eur Heart J, 33, 165–175.

111.

Voora

, and Ginsburg

. (2012). Clinical application of cardiovascular pharmacogenetics. J Am Coll Cardiol, 60, 9–20.

112.

Yang

, Lai

, Parnell

, et al. (2005). Genome-wide linkage analyses and candidate gene fine mapping for HDL3 cholesterol: the Framingham study. J Lipid Res, 46, 1416–1425.

113.

Zekraoui

, Lagarde

, Raisonnier

, Gerard

, Aouizerate

, and Lucotte

. (1997). High frequency of the apolipoprotein E *4 allele in African pygmies and most of the African populations in sub-Saharan Africa. Hum Biol, 69, 575–581.

114.

Zhang

, Yu

, He

, et al. (2006). Role of BCRP 421C>A polymorphism on rosuvastatin pharmacokinetics in healthy Chinese males. Clin Chim Acta, 373, 99–103.

115.

Zhao

, Tuakli-Wosornu

, Lagace

, et al. (2006). Molecular characterization of loss-of-function mutations in PCSK9 and identification of a compound heterozygote. Am J Hum Genet, 79, 514–523.