The PON1 M55L gene polymorphism is associated with reduced HDL-associated PAF-AH activity.

The platelet-activating factor acetylhydrolase activity associated with high density lipoprotein (HDL-PAF-AH) may substantially contribute to the antioxidant, anti-inflammatory, and overall antiatherogenic effects of HDL. Two enzymes associated with HDL express PAF-AH catalytic activity, PAF-AH itself and paraoxonase-1 (PON1). The relative contribution of these enzymes in the expression of PAF-AH activity on HDL remains to be established. We investigated whether the PON1 polymorphisms (M55L and Q192R) or the PAF-AH polymorphism V379A could affect the PAF-AH activity associated with HDL in both normolipidemic and dyslipidemic (type IIA and IIB) populations. We show for the first time that the PON1 M55L polymorphism significantly affects the HDL-PAF-AH activity in all studied groups, the PON1 L55L individuals having lower enzyme activity compared to those having 1 M and 2 M alleles. No differences in the HDL content concerning the major apolipoprotein and lipid constituents were observed between individuals carrying the PON1 L55L and those with the M55M polymorphism. Our results provide evidence that PON1 significantly contributes to the pool of HDL-PAF-AH activity in human plasma, and suggest that the low PAF-AH activity in HDL carrying the PON1 L alloenzyme may be an important factor contributing to the low efficiency of this HDL in protecting LDL against lipid peroxidation.

It is well established that HDL plays a protective role against atherogenesis and coronary heart disease (CHD) (1). This effect appears to be multifactorial and involves a number of mechanisms, including reverse cholesterol transport, as well as antiinflammatory and antioxidant activities. Potential roles in the HDL antiinflammatory and antioxidant properties are played by two enzymes associated with this lipoprotein, platelet-activating factor acetylhydrolase (PAF-AH) and paraoxonase-1 (PON1) (2).
PAF-AH exhibits a ␣ / ␤ hydrolase conformation and has broad substrate specificity toward lipid esters containing short acyl chains. It also displays phospholipase A 1 and A 2 activities as well as transacetylase activity. Of these, the Ca 2 ϩ -independent phospholipase A 2 activity of PAF-AH has been principally studied [reviewed in ref. (3)]. PAF-AH can hydrolyze and inactivate PAF, the potent lipid mediator involved in inflammatory diseases as well as in atherogenesis. PAF-AH can also effectively hydrolyze oxidized phospholipids that play key roles in several aspects of atherogenesis. In human plasma, PAF-AH is primarily associated with LDL particles, whereas a small proportion ( Ͻ 20% of total enzyme activity) is associated with HDL (3). The gene for the plasmatic PAF-AH has been mapped to chromosome 6q21.2-p12, comprises 12 exons, and spans at least 45 kb of DNA (4). Several genetic variants of this enzyme have been described. Two of them are situated in exon 9 (V279F and Q281R) and have been described only in Japanese populations. These mutations result in a complete loss of the enzyme activity and are associated with severe asthma, stroke, and cardiovascular disease (5,6). Two other genetic variants, V379A (exon 11) and I198T (exon7), have been described and are associated with atopy and asthma in European populations. Furthermore, in vitro studies have shown that these two polymorphisms are associated with a reduction in PAF-AH kinetic constants (7).
PON1 is an esterase that in plasma is exclusively associated with HDL. PON1 can hydrolyze oxidized phospholipids and cholesteryl ester hydroperoxides formed dur-ing lipoprotein peroxidation, and plays a protective role against the oxidative modification of plasma lipoproteins (8). The PON1 gene is located on the long arm of chromosome 7. Two genetic variants of this enzyme have been described, one at position 55 (methionine/leucine, M55L) and the other at position 192 (glutamine/arginine, Q192R). The substrates for PON1 that have been mostly used in assays in vitro are paraoxon and phenyl acetate. Several studies have shown that paraoxon hydrolytic activity using either HDL or purified PON1 is high in PON1 R192R and L55L individuals and low in PON1 Q192Q and M55M individuals. By contrast, both PON1 alloenzymes have a similar hydrolytic activity toward phenyl acetate (9). On the other hand, the capacity of PON1 alloenzymes to protect LDL from oxidation is distinguished from that of paraoxon hydrolytic activity. Thus, PON1 either purified or associated with HDL from MM/QQ individuals exhibits the greatest protective capacity toward LDL oxidation in vitro (10).
In addition to the above activities, PON1 expresses a PAF-AH catalytic activity (11). In this context, it has been suggested that the HDL-associated PAF-AH (HDL-PAF-AH) activity is mainly due to the PAF-AH catalytic activity of PON1, as no PAF-AH protein was found on this lipoprotein (11). By contrast, it was recently shown that the total PAF-AH activity associated with HDL is due to the PAF-AH protein (12). In the present study we investigated whether the PON1 polymorphisms (M55L and Q192R) or the PAF-AH polymorphism V379A could affect the PAF-AH activity in plasma and that associated with HDL in patients with dyslipidemia of type IIA or IIB, as well as in normolipidemic, apparently healthy volunteers.

Patients
Two hundred and two unrelated dyslipidemic patients attending the Lipid Clinic of the University Hospital of Ioannina participated in the study. After the initial screening, patients were advised to follow the National Cholesterol Education Program step 1 diet (which limits dietary intake of cholesterol to 300 mg/day, that of saturated fats to 10% of total energy intake and total fats to 30% of total energy intake) for 3 months. At the end of this period, a complete laboratory baseline analysis was performed. According to their lipid levels, patients were divided into the primary hypercholesterolemia (dyslipidemia type IIA) group (LDL cholesterol Ͼ 160 mg/dl) and the combined hyperlipidemia (dyslipidemia type IIB) group (LDL cholesterol Ͼ 160 mg/dl and triglycerides Ͼ 200 mg/dl). Among dyslipidemic IIA patients, 58 had a diagnosis of heterozygous familial hypercholesterolemia and 80 had a diagnosis of polygenic hypercholesterolemia.
A total of 122 healthy unrelated volunteers with no history of dyslipidemia or CHD also participated in the study. Volunteers were either staff of the University Hospital of Ioannina or attending for a routine health check at the Outpatient Clinic of this Hospital.
None of the participants were taking lipid-lowering drugs or any other medication known to affect lipid metabolism, including hormonal therapy, during the last 12 weeks. Individuals with hypertension (blood pressure, 140/90 mmHg on repeated measurements), diabetes mellitus (fasting blood glucose, 126 mg/dl), obesity (body mass index, 30 kg/m 2 ) or thyroid, hepatic, or renal diseases, as well as subjects known to ingest more than two alcoholic drinks daily or who were taking vitamin supplements were excluded from the study. Finally, no patient had any clinical evidence of cardiovascular disease. The study was approved by the Ethics Committee of the University Hospital of Ioannina, and all study patients gave written informed consent for participation in the study.

Determination of PON1 and PAF-AH genotypes
Genomic DNA was obtained from leukocytes using standard procedures. The PON1 M55L and Q192R polymorphisms were detected using a previously reported protocol (13). To determine the variations of 279 (14) and 379 (7) of PAF-AH genes, DNA was amplified and analyzed by restriction isotyping. Briefly, primers for amplification of a 99 bp DNA that contains the coding sequence for position 192 were 5 Ј TAT TGT TGC TGT GGG ACC TGA G3 Ј and 5 Ј CAC GCT AAA CCC AAA TAC ATC TC3 Ј . After an initial denaturation step of 5 min at 95 Њ C, the PCR was carried out for 40 cycles, with each cycle consisting of 60 s of denaturation at 94 Њ C, 45 s of annealing at 56 Њ C, and 45 s of extension at 72 Њ C. PCR product was digested with 5 units of Alw I restriction enzyme for 3 h at 37 Њ C.
For genotyping the M55L polymorphism, the primers for amplification of 144 bp DNA encoding codon 55 were 5 Ј GAG TGA TGT ATA GCC CCA GTT TC3 Ј and 5 Ј AGT CCA TTA GGC AGT ATC TCC G3 Ј . An initial incubation for 5 min at 95 Њ C was followed by the step of amplification that was carried out for 40 cycles, with each cycle consisting of 1 min of denaturation at 94 Њ C, 45 s of annealing at 61 Њ C, and 45 s of extension at 72 Њ C. PCR product was digested with 5 units of Hinf I restriction enzyme for 24 h at 37 Њ C.
The PAF-AH V379A polymorphism was detected using 5 Ј GGG AGA CAT AGA TTC AAC TG3 Ј and 5 Ј GGT CAT GAA AAA AAT AGT TT3 Ј primers. After an initial denaturation step of 5 min at 94 Њ C, the PCR was carried out for 35 cycles, with each cycle consisting of 45 s of denaturation at 94 Њ C, 50 s of annealing at 53 Њ C, and 50 s of extension at 72 Њ C. PCR product was digested with 5 units of Pst I restriction enzyme for 24 h at 37 Њ C.

Isolation of HDL
HDL (d ϭ 1.063-1.210 g/ml) was isolated from the HDL-containing supernatant after treatment of serum with magnesium chloride/dextran sulfate to precipitate all apolipoprotein B (apoB)-containing lipoproteins (HDL-rich serum). HDL isolation was performed by sequential ultracentrifugation in a Beckman L7-65 ultracentrifuge at 40,000 rpm, 14 Њ C, with a type NVT 65 rotor. The HDL preparation was dialyzed against 10 mM PBS (pH 7.4) for 24 h at 4 Њ C. It was then filter sterilized and stored in the dark at 4 Њ C under nitrogen for up to 2 weeks.

Measurement of PAF-AH and PON1 activities
PAF-AH activity in isolated HDL, in plasma, and in HDL-rich plasma (prepared as the HDL-rich serum) was determined by the trichloroacetic acid precipitation procedure using [ 3 H]PAF (100 mM final concentration) as a substrate (15). PON1 activity in serum and in HDL-rich serum was measured using paraoxon as a substrate, whereas PON1 activity in isolated HDL was measured using phenyl acetate as a substrate. Both PON1 activities were determined in the presence of 2 mM Ca ϩ 2 in 100 mM Tris-HCl buffer (pH 8.0) for paraoxon and in 20 mM Tris-HCl buffer (pH 8.0) for phenyl acetate (15).

Analytical methods
Serum total cholesterol and triglycerides were determined on the Olympus AU560 Clinical Chemistry analyzer (Hamburg, Ger-by guest, on July 19, 2018 www.jlr.org Downloaded from many). Serum HDL cholesterol levels were measured with the above method in the HDL-rich serum. Serum LDL cholesterol levels were calculated using the Friedewald formula. Serum apoB, apoA-I, apoA-II, and apoE were measured by immunonephelometry with the aid of a Behring Nephelometer BN100 and reagents (antibodies and calibrators) from Behring Diagnostics GmbH (Liederbach, Germany). The total cholesterol and the phospholipid content of HDL were measured enzymatically using the Bio-Merieux kit (15).

Statistical analysis
One-way ANOVA followed by least significant difference test was used for comparisons among individual groups. Allele frequencies were estimated by the gene-counting method, and Hardy-Weinberg's equilibrium was tested by the 2 test. The same test was used to compare percentages. Mann-Whitney U test and Kruskal-Wallis test were used for comparisons among the paraoxonase activities. Student's t -test for independent samples was used to estimate the effect of the enzyme polymorphisms on lipid parameters and enzyme activities.

Clinical and biochemical characteristics of the study population
The clinical and biochemical characteristics of the populations participating in the study are shown in Table 1 . As expected, the dyslipidemic populations had significantly higher levels of all lipidemic parameters compared with normolipidemic individuals, with the exception of HDL cholesterol levels in type IIB patients, which were lower compared with other groups. Plasma PAF-AH activity was higher in type IIA and IIB patients compared with controls, but there was no difference in enzyme activity between the two dyslipidemic populations. In accordance with our previously published results (15), the HDL-PAF-AH activity was significantly lower in type IIB patients com-pared with the other studied groups ( P Ͻ 0.001). Finally, no difference in serum PON1 activity toward paraoxon measured either in total serum or in HDL-rich serum (HDL-PON1) was observed among the studied groups (Table 1).

PON1 and PAF-AH genotyping analysis
The genotype and allele frequencies studied are shown in Table 2 . Importantly, no difference in the distributions of both PON1 polymorphisms as well as in the allele frequencies was observed among the studied groups. Combined haplotypes (both haplotypes per subject) were constructed for PON1 polymorphisms and confirmed the well-known linkage disequilibria between M55L and R192Q ( Table 2). All 2 tests to compare observed and expected genotype frequencies have been performed and all were in Hardy-Weinberg equilibrium.
No difference in the distributions of the PAF-AH V379A polymorphism and in the allele frequencies was observed among the studied groups. As expected, no one from the study populations exhibited the PAF-AH V279F mutation, which has been found only in the Japanese population (14).

Influence of PON1 and PAF-AH polymorphisms on lipid profile and enzyme activities
Individuals of each group were subgrouped as follows: i ) according to PON1 192 genotype into homozygotes for the Q allele and those having one or two R alleles; ii ) according to PON1 55 genotype into homozygotes for the L allele and those having one or two M alleles; iii ) according to PAF-AH 379 genotype into homozygotes for the A allele and those having one or two V alleles. Comparisons between the above subgroups were performed within each studied group. These subgroups were matched for age, sex, body mass index, and smoking habits.
No difference in the serum lipid and apolipoprotein levels was observed among the above genotype subgroups (data not shown), suggesting that none of the studied polymorphisms affect the lipid profile either in healthy volunteers or in any patient group. As shown in Table 3, the PON1 Q192R polymorphism significantly influenced the serum PON1 activity toward paraoxon in each studied group, the Q192Q individuals having lower activity compared with the R carriers of the same group, an observation that accords with previously published results (16). The same phenomenon was observed when subjects of each group were subgrouped according to the PAF-AH V379A polymorphism (Table 3). It must be noted that when the R carriers were divided into homozygotes and heterozygotes, the Q192R individuals presented intermediate levels of enzyme activity between the Q and R homozygotes of the same group (data not shown). Furthermore, independent of the presence of the PAF-AH V379A polymorphism, the PON1 activity was significantly influenced by the M55L polymorphism, the L55L individuals having higher activity compared with the M-carriers in each studied group (Table 3), a finding that also accords with previously published data (16). When the M carriers were divided into homozygotes and heterozygotes, the M55L individuals presented intermediate levels of enzyme activity between the M and L homozygotes of the same group (data not shown). It must be noted that similar results were obtained for the PON1 activity measured in the HDL-rich serum (data not shown). The present study shows for the first time that neither the Q192R nor the M55L polymorphism of PON1 influences the plasma PAF-AH activity. Furthermore, the plasma PAF-AH activity was not influenced by the PAF-AH V379A polymorphism (Table 3).
The most important finding of the present study is that the PON1 M55L polymorphism significantly affects the HDL-PAF-AH activity. As shown in Table 3, the PON1 LL homozygotes of each group exhibited significantly lower HDL-PAF-AH activity compared with M carriers of the same group. The same phenomenon was observed when subjects of each group were subgrouped according to the PAF-AH V379A polymorphism (Table 3). When the M carriers were divided into homozygotes and heterozygotes, the M55L individuals presented intermediate levels of enzyme activity between the M and L homozygotes of the same group (HDL-PAF-AH activity in nmol/ml/min; normolipidemic group: M55L ϭ 3.2 Ϯ 0.8 and M55M ϭ 3.9 Ϯ 0.9, P Ͻ 0.05; type IIA: M55L ϭ 3.3 Ϯ 0.7 and M55M ϭ 4.0 Ϯ 0.9, P Ͻ 0.05; type IIB: M55L ϭ 2.5 Ϯ 0.6 and M55M ϭ 3.2 Ϯ 0.7, P Ͻ 0.05). The above results are further supported by those obtained when the comparisons of the HDL-PAF-AH activity in each group were performed according to the PON1 M55L and Q192R haplotypes. Thus, in each studied group, subjects with the L allele had lower enzyme activity compared with those having the M allele independent of the existence of the Q or R allele. Figure 1 presents the results for the type IIA dyslipidemic patients. Finally, there was no influence of the PON1 Q192R or the PAF-AH V379A polymorphism on the HDL-PAF-AH activity.

Compositional analysis of HDL
In an effort to provide a mechanistic explanation for the lower PAF-AH activity in HDL of subjects carrying the PON1 L55L polymorphism, we performed compositional analysis of HDL isolated from the HDL-rich serum of 10 subjects from each group carrying the PON1 L55L or the PON1 M55M polymorphism. ApoA-I, apoA-II, and apoE, the phospholipids, and the total cholesterol content of HDL, as well as the PAF-AH activity and the PON1 activity toward phenyl acetate, were determined. As shown in Table 4, the PAF-AH activity of the HDL isolated from the L55L individuals was significantly lower compared with the enzyme activity of the M55M individuals of the same group. By contrast, no differences in the other parameters of HDL were observed between individuals with the PON1 L55L and those with the PON1 M55M polymorphism within each studied group.

DISCUSSION
Several lines of evidence suggest that the HDL-PAF-AH activity may substantially contribute to the antioxidant and antiinflammatory effects of HDL, thus this activity may be an important component of the multiple mechanisms by which HDL slows the progression of atherosclerosis (17). The HDL-PAF-AH activity may represent a pool of similar catalytic activities expressed by three different enzymes, PAF-AH, PON1, and LCAT (3,17). However, the relative   (18). However, according to our results, no alterations in the lipid composition of HDL or in the apoA-II/apoA-I ratio were observed between L55L and M55M individuals in any studied group. Furthermore, it has been shown that among the HDL particle populations, PAF-AH is mainly associated with the apoErich HDL particles (19). Thus, alterations in the apoE content of HDL might have influenced the association of PAF-AH with this lipoprotein. However, according to our results, there was not any difference in the HDL apoE/apoA-I ratio between L55L and M55M individuals. Consequently,  due to the lack of differences in the HDL composition between L55L and M55M individuals, we may exclude the possibility that the low enzyme activity in HDL from carriers of the L allele can be attributed to any influence of the HDL composition either on the PAF-AH activity or on the enzyme affinity to HDL particles in this population. It has been shown that PON1 exerts a PAF-AH activity (11), and at the same time, this enzyme is capable of oxidatively hydrolyzing fragmented phospholipids (20). These activities confer on PON1 the ability to retard LDL oxidation (21) and to reduce the proinflammatory effects of oxidized LDL (8). These PON1 activities are calcium independent and are distinguished from the PON1 hydrolytic activity against organophosphate substrates, which are highly calcium dependent (22). The contribution of PON1 to the PAF-AH activity expressed on HDL has recently been a subject for debate. Thus, it has been suggested that the HDL-PAF-AH activity is primarily due to PON1, although it was recently shown that the PAF-AH activity in HDL is exclusively due to the PAF-AH protein (12). Our finding that the PON1 M55L polymorphism significantly influences the HDL-PAF-AH activity rather favors the assumption that PON1 essentially contributes to the HDL-PAF-AH activity. We may exclude the possibility that the HDL of L carriers contains less PON1 protein compared with the M carriers, as we did not find any difference between L55L and M55M individuals in the HDLassociated PON1 activity toward phenyl acetate, which is not influenced by this polymorphism (9). Moreover, it has been shown that the PON1 allele is associated with higher serum concentrations of the enzyme compared with the M allele (23). Based on the above observations, we may suggest that the low PAF-AH activity in individuals carrying the L alloenzyme may be due to the replacement of 55-methionine by leucine, thus suggesting that the 55methionine residue of PON1 is essential for the expression of the PAF-AH catalytic activity by this enzyme.
As expected, none of the subjects investigated exhibited the loss-of-function PAF-AH V279F mutation. On the other hand, one-third of our population exhibited the PAF-AH V379A polymorphism, which, however, did not affect either total plasma PAF-AH activity or the HDL-associated enzyme activity. It has been previously shown that this polymorphism significantly affects the enzyme kinetic constants (7); however, these experiments were performed using purified recombinant enzyme carrying the V379A variant. Thus, the difference between these previously published results and those of the present study could be due to other factors that influence the enzyme activity, such as the lipid environment of the lipoprotein particles that carry the enzyme in plasma (18).
The lower PAF-AH activity in L55L individuals compared with the carriers of the M allele found in the present study may explain, at least partially, the lower capacity of PON1 from the L55L individuals to protect LDL from oxidation compared with the carriers of the M allele (24). However, based on our results, we cannot comment on whether the low HDL-PAF-AH activity in L55L individuals may contribute to the association between the PON1 55L allele and athero-sclerosis reported in previous studies (25,26). On the other hand, other studies failed to show any association of this allele with the risk of CHD (27). Additionally, the PON1 Q192R polymorphism has been associated with the risk for CHD (28), and according to our results, this polymorphism does not affect the HDL-PAF-AH activity. Thus, it is profound that further studies involving CHD patients are required to show whether the results of the present study have any impact on the pathophysiology of CHD.
In conclusion, the present study shows for the first time that the PON1 M55L polymorphism influences the PAF-AH activity associated with HDL in both dyslipidemic and normolipidemic populations, and further suggests that PON1 significantly contributes to the pool of PAF-AH activity associated with HDL in human plasma.