Each year more than 2 million individuals undergo coronary stenting¹, a procedure in which a bare or drug-coated metal stent is placed in an atherosclerotic plaque to improve myocardial blood flow. This foreign body deployment in an arterial circuit subject to high shear stress increases risk of platelet-mediated thrombosis. To counter the chance of such an occurrence, people subjected to this procedure receive a combination antiplatelet therapy with aspirin and clopidogrel. This routine adjunct pharmacologic regimen, for one of the most commonly performed medical procedures, has contributed to making clopidogrel the second most prescribed drug in the world, with over $9 billion in sales in 2010 (ref. 2).

Clopidogrel is a prodrug that is metabolized to its active form in vivo and then inhibits the ADP P2Y12 receptor, a key mediator of platelet aggregation. For many years, it has been recognized that the platelet inhibitory response to clopidogrel is heterogeneous³. This variable response is particularly concerning among people who have had a stent placed, as clotting of the metal sleeve usually results in sudden death or a heart attack. Whereas the incidence of stent thrombosis is relatively low (1–2%) with the current dual antiplatelet therapy, it represents a catastrophic complication for thousands of people every year⁴. Elucidating why clopidogrel is inconsistent in suppressing platelet aggregation is therefore crucial.

In this issue of Nature Medicine, Bouman et al.⁵ found a new mediator of clopidogrel's platelet effect, PON1, that drives the conversion of the drug into the active metabolite. In humans, a functional polymorphism of PON1, PON1 Q192R, affects the platelet response, clopidogrel pharmacokinetics and the risk for thrombosis. PON1 QQ192 homozygous individuals show limited platelet inhibition and decreased plasma amounts of both active PON1 and active clopidogrel metabolite. Therefore, genotyping of PON1 might be used to identify clopidogrel responders and help predict clinical efficacy.

Insights into clopidogrel heterogeneity began with a candidate gene study of platelet function in healthy volunteers that pointed to a common polymorphism in the cytochrome 2C19 gene (CYP2C19), which had been implicated in the metabolic activation of clopridogrel⁶. In 2008, three large clinical studies in people who had stent procedures validated the association of loss-of-function CYP2C19 alleles with more than threefold higher risk of stent thrombosis^{7, 8, 9}. More recently, two large meta-analysis studies of ten independent clinical trials have confirmed this increased risk associated with loss-of-function CYP2C19 alleles^{10, 11}.

Indeed, using the quantitative trait of platelet response to clopidogrel as the phenotype, the unbiased approach of genome-wide association testing identified the cytochrome cluster, specifically CYP2C19, as the only common genomic variant to account for the heterogeneous clopidogrel effect¹². But, as with the results from most genome-wide association studies, the variants in the gene only accounted for a small fraction (~12%) of the overall variability of the clopidogrel platelet response.

Using in vitro liver microsomal assays and separating the bioactivation of clopidogrel into two main steps (Fig. 1), the authors found that PON1, an esterase associated with high-density lipoprotein concentrations in the blood, emerged as the rate-limiting enzyme for the second step of hydrolytic cleavage of the 2-oxo-clopidogrel ring to the thiol active metabolite.

The first is modulated by cytochromes and converts clopidogrel to 2-oxo-clopidogrel, which is still not active. The second step involves PON1-mediated hydrolytic cleavage of the 2-oxo-clopidogrel ring structure to a thiol and yields the fully active metabolite. As a consequence of diminished conversion of clopidogrel to the thiol metabolite by the PON1 Q192 allele, there is an increased risk of stent thrombosis, a platelet-mediated event.

Marina Corral

• Full size image (125 KB)

Of the two previously identified common, nonsynonymous coding variants of PON1, the Q192R variant was associated with stent thrombosis in a cohort of people with coronary artery disease⁵. The association was substantial with QR192 heterozygotes having more than fourfold risk for thrombotic episodes and more than 12-fold risk for QQ192 homozygotes. A second prospective cohort of people with acute coronary syndromes replicated this association with more than three- and tenfold risk of developing stent thrombosis among heterozygotes and homozygotes, respectively.

Bouman et al.⁵ also found considerable biochemical data to corroborate their PON1 findings, including increased amounts of the inactive drug 2-oxo-clopidogrel, decreased levels of the thiol active metabolite and less platelet inhibitory effect among people with stent thrombosis carrying a PON1 Q192 allele. Furthermore, the significantly lower (by 64%, P = 0.006) bleeding rate among QQ192 homozygotes compared with RR192 individuals supports the reduced platelet effect of clopidogrel for individuals with this common variant. Overall, the PON1 Q192R variant accounted for most (over 70%) of the variability in response to platelet aggregation after clopidogrel treatment.

Characterizing the effect of the PON1 polymorphism is clearly a crucial step forward in understanding clopidogrel pharmacogenomics. The risk of stent thrombosis in heterozygous PON1 QR192 individuals is at least as high, and might be substantially greater in QQ192 homozygotes, compared to individuals with CYP2C19 loss-of-function alleles⁵.

However, there are some difficult points to reconcile. At the metabolomic level, Bouman et al.⁵ did not find CYP2C19 to be active in the initial oxidation of clopidogrel. Variations in CYP2C19 were also not associated with stent thrombosis in the two cohorts that they assessed. Moreover, the common PON1 variant was not identified in the only genome-wide association study of clopidogrel's impact on platelet aggregation¹². Whereas new genetic polymorphisms would be expected to be associated with clopidogrel's antiplatelet effect, these contradictions might not have been anticipated.

Bouman et al.⁵ offer some possible explanations for these contradictions. The only genome-wide association for clopidogrel pharmacology was performed in Amish individuals, an inbred restricted population, and the phenotype was assessed 7 d after treatment, once steady-state clopidogrel effects would be anticipated. In the current study, the levels of the active metabolite and platelet effect were assessed right after a single dose of clopidogrel was administered in a more diverse cohort of European ancestry. Nonetheless, it remains unclear why the PON1 study did not replicate the considerable body of data validating the clinical impact of CYP2C19 loss-of-function alleles.

Notwithstanding this issue, this study has vital implications and suggests next steps in research and clinical practice. We desperately need a hypothesis-free and comprehensive assessment of clopidogrel pharmacogenomics. This PON1 study was a candidate gene study supported by in vitro metabolic experiments that nominated this gene and independently replicated the association in a second cohort. Not only might there be other important, as yet undiscovered PON1 variants, but also there are probably other gene polymorphisms associated with biological properties of clopidogrel other than its activation, such as its absorption from the gut or its binding of the P2Y12 receptor. Full exome or whole-genome sequencing approaches could yield such information on both platelet responsiveness and clinical outcomes with clopidogrel therapy.

As the risks of thrombosis for people after coronary stenting are high, and genotyping can be accomplished quite inexpensively, PON1 and CYP2C19 genotyping should certainly be considered in the clinic¹³. This practice is especially bolstered by providing actionable information for selection of other P2Y12 inhibitors, such as prasugrel or ticagrelor, which are unlikely to be influenced by polymorphisms in either of these genes. Ultimately, a full panel of functional genomic variants that comprehensively addresses the pharmacogenomics of clopidogrel will become part of standard clinical care for people who undergo stenting procedures.