Paraoxonase-1 is an atherosclerosis protective enzyme associated with HDL and systemic anti-oxidant function. Its catalytic activity within crude serum mixtures has traditionally been measured by quantifying enzymatic hydrolysis rates of two known in vitro substrates, with functional activities named paraoxonase and arylesterase activities4, 5
. There are several novel findings in this report that describe the clinical and genetic associations of PON-1 activities with cardiovascular risks that differ from prior observations from smaller cohorts. First, we demonstrated prognostic value of serum arylesterase activity (and to a lesser extent serum paraoxonase activity) in predicting long-term cardiovascular risk in a wide range of subjects already treated with contemporary medical therapy. We observed that diminished serum arylesterase activity, particularly within the lowest quartile range, was predictive of adverse long-term cardiac events independent of standard clinical and biochemical risk factors, and provided incremental value in reclassifying subjects who are at higher risk of long-term MACE. Furthermore, the prognostic value of serum arylesterase activity was evident in both primary as well as secondary prevention subjects. The fact that arylesterase activity provided prognostic value consistently within subjects with or without underlying angiographic evidence of significant CAD implies that a lack of systemic anti-oxidant defense mechanisms (a primary function of PON-1) may both promote greater vulnerability to oxidative stress, as well as increase risk for development and progression of CAD in subjects. Moreover, the ability of low serum arylesterase activity to identify those at significant increased risk of MACE, even among primary prevention subjects with recent coronary angiographic data showing < 50% stenosis in all major coronary vessels, suggests this assay has prognostic value, and can identify a vulnerable cohort of subjects who otherwise are not identified as being at high risk. Taken together, the present results suggest that enzymatic activity measures with arylesterase, more so than paraoxonase, serve as a powerful prognostic indicator of cardiovascular risk in a broad spectrum of subjects.
Studies on PON-1 activities have relied on quantifying its wide ranges of enzymatic activities in breaking down in vitro substrates like paraoxon (paraoxonase activity) and phenylacetate (arylesterase activity). These activities are often reported together and the findings are commonly concordant. Hence, the distinction between serum paraoxonase and arylesterase activities in predicting future adverse cardiac events is somewhat unexpected, even though historically the correlation between these two measures has not been particularly tight. The unique relationship between the two PON-1 activity measurements is largely the result of a very strong association between serum paraoxonase activity and its underlying genetic determinants, which may also explain why serum arylesterase activity is normally distributed while serum paraoxonase activity was not. Meanwhile, the strong association between the PON gene cluster and serum arylesterase activity confirms the long-standing assumption that this locus harbors important genetic determinants of serum arylesterase activity.
The significant genetic associations revealed by the GWAS analyses for paraoxonase and arylesterase activities are of further interest for several reasons. First, there was a stronger relationship between serum paraoxonase activity levels and its genetic determinants than those with serum arylesterase activity. Second, our results confirm the strong association of PON-1 variants with serum paraoxonase and arylesterase activities and demonstrate that the two measured enzymatic functions are associated in a Mendelian-like fashion by a single, major locus on chromosome 7 containing the PON gene cluster. This is supported by the GWAS analyses conditioned on the lead SNPs, which did not identify other genomic regions associated with either paraoxonase or arylesterase activity. Furthermore, the lead SNPs for either enzymatic activity localize to the PON-1 gene, are distinct from each other, and are not in LD with SNPs in the neighboring PON-3 and PON-2 genes. However, while the PON-1 variants exhibited strong (if not opposed) effects on PON function, they were not associated with future risk of MACE, either individually or as a haplotype that modestly decreased both paraoxonase and arylesterase activity. These results are consistent with our analyses from the CARDIoGRAM consortium, which did not reveal an association of these SNPs with either prevalent CAD or history of MI in ~80,000 subjects. It should be noted, however, that since the lead SNPs only contributed to an estimated ~15% of PON-1 activity variation, other processes, such as post-translational modifications, could also play a role in determining the ultimate functionality of PON-1. For example, PON1 is known to be sensitive to post-translational oxidative modification and inactivation14
. It is also possible that our study was underpowered to detect genetic effects on prospective risk of MACE since only 311 subjects out of the ~2,700 subjects included in these analyses experienced a MACE over three years of follow up.
The lack of associations between the lead SNPs for PON and arylesterase activities with CAD and MI risks in humans in our study are in direct contrast to what has been observed in mice where PON-1 deficiency leads to increased aortic lesion formation3
, and transgenic mice over expressing PON-1 are protected 15, 16
. It is of interest that the PON-1 transgenic mouse models demonstrating protective effects have 50–400% increased serum arylesterase activity 15, 16
, far in excess of the modest changes in activity associated with the peak SNP identified in the arylesterase GWAS. If arylesterase activity is the more important atheroprotective aspect of PON-1 function, as our clinical associations with MACE suggest, then it may not be surprising that an association was not observed with the SNPs identified. For example, the lead SNP for arylesterase activity (rs854572) only increases activity by ~16% per minor allele copy. It is thus possible that the genetic effects on arylesterase activity from this SNP are too weak to observe, especially if a minimum biological threshold of activity change is needed to influence risk of prevalent CAD, history of MI, or incident risk of MACE.
The ability for serum arylesterase activity to identify a high-risk population in the primary prevention cohort that just underwent cardiac catheterization and showed no significant evidence of stenoses in any major vessel may have important clinical implications. Even though our subset analysis was limited by the relatively smaller sample size and low event rates, a significant 4-fold increase risk in long-term MACE in these otherwise "low risk" subjects was identified. Use of serum arylesterase activity may thus have clinical utility, and help identify an important patient cohort that may warrant more aggressive risk factor reduction treatment strategies who might otherwise not be targeted for aggressive preventive intervention. In addition to low serum arylesterase activity predicting increased risk for MACE in primary prevention subjects, there was a direct association also noted between lower serum arylesterase activity level and increased prevalence of significantly obstructive CAD by angiography. Of note, we did not find a strong relationship between systemic inflammatory biomarkers, such as C-reactive protein or leukocyte parameters, with serum arylesterase levels in this group of stable cardiac patients. This observation is consistent with prior reports that distinguished systemic inflammatory from oxidative stress processes.
Several limitations of our study should be noted. Serum paraoxonase and arylesterase activities utilized substrates that are not the endogenous substrates for PON-1, but are used because they are not readily influenced by other esterases/lactonases in serum, and are presumed to reflect the underlying catalytic activity of PON-1. The GWAS results observed demonstrate these assumptions are reasonable, as only genetic variations in the PON-1 gene were observed to be associated with variations in paraoxonase and arylesterase activities. A further potential limitation is that the measurements were only made under fasting conditions at a single time-point. Hence, we are unable to determine the variability and prognostic value of level changes over time and the impact of dietary or therapeutic interventions on serum arylesterase activity level. Selection bias may also be present for those undergoing cardiac catheterization for symptomatic evaluation and management of cardiac diseases at a tertiary care setting, but the large sample size and event rates of the patient population, together with careful phenotypic evaluation including angiographic data, provides unique insights and adequate power to adjust for clinical and biomarker variables.