INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Benveniste and coworkers (1) found a substance released from rabbit basophils that activates platelets, calling it platelet-activating factor (PAF). PAF (1-o-alkyl-2-acetyl-sn-glycero-3-phosphocholine) mediates various pathophysiological functions in the inflamed airway such as chemotaxis and activation of inflammatory cells, increased vascular permeability and mucosal edema, and prolonged airway hyperresponsiveness (2). Overexpressed PAF receptors in transgenic mice induced enhanced airway reactivity to methacholine (3), while PAF receptor antagonists reduced airway hyperresponsiveness in allergen-sensitized guinea pigs or patients with asthma. Despite this evidence, it is still controversial whether the PAF concentration in bronchial lavage fluid increases after allergen provocation in atopic asthmatics (4). Interestingly, the concentration of lyso-PAF, a biologically inactive metabolite of PAF, rose by more than tenfold in the allergen-exposed nasal mucosa and airway (4, 8), suggesting that PAF is immediately degraded to lyso-PAF once released in the airway.

PAF acetylhydrolase (PAF-AH, E.C.3.1.1.47) hydrolyzes PAF to lyso-PAF by removing the acetyl group at the sn-2 position of glycerol backbone. Two features differentiate PAF-AH from other phospholipase A₂; first, the enzyme activity is Ca²⁺-independent; second, PAF-AH demonstrates a substrate specificity for the short and/or oxidized acyl group at the sn-2 position of glycerol (9). Among several PAF-AH isoforms identified so far, plasma PAF-AH is the only extracellular one. Tjoelker and colleagues (10) found that either locally or systematically administered recombinant PAF-AH blocked PAF-induced paw edema or pleurisy by > 80% in mice, suggesting the ability of this enzyme to break down PAF at the site of inflammation. Plasma PAF-AH concentrations vary in physiological and pathological conditions such as pregnancy, parturition, hyperlipidemia, cardiovascular diseases, stroke, rheumatoid arthritis, sepsis, and asthma (11). About 4% of the Japanese population is known to lack plasma PAF-AH activity completely (18). The molecular mechanism of plasma PAF-AH deficiency in Japanese was identified by Stafforini and colleagues (19) as a missense mutation (G T mutation at 994th nucleotide) in exon 9 of the plasma PAF-AH gene; this mutation replaces a valine residue at position 279 with a phenylalanine (V279F) causing enzyme activity loss and impaired protein transport.

We hypothesized that plasma PAF-AH deficiency caused by V279F mutation may modify asthma prevalence and phenotypes in the Japanese population. We analyzed plasma PAF-AH genotype and/or activity in 279 unrelated Japanese asthmatic adults and 217 healthy subjects to determine the association of V279F mutation with asthma prevalence, type, and severity.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects

Some 279 consecutive asthma patients visiting Keio University Hospital between December 1996 and May 1997 were enrolled in this study (152 male, 127 female, age 15 to 87 yr). All patients met the Guidelines for the Diagnosis and Management of Asthma from the National Institutes of Health (20). Patients with a recent episode of acute exacerbation were excluded until respiratory symptoms had stabilized more than 4 wk. Patient profiles, current status of symptoms and medication, pulmonary function test data, total and/or allergen-specific serum IgE levels were obtained from charts and written questionnaires. Patients were identified as atopic based on elevated serum IgE level and/or evidence of a specific IgE antibody to one or more of the following antigens: Der p1, Der f2, and Fel d1. Severity of asthma for each patient was determined based on the NIH Guidelines (20). We also recruited 217 healthy volunteers without history of atopic diseases or asthma (111 male, 106 female). The protocol was approved by the institutional review board of Keio University Hospital.

Plasma PAF-AH Gene Genotyping

Blood was drawn by venipuncture in a tube containing ethylenediaminetetraacetic acid (EDTA) 2Na, centrifuged at 1,500 × g for 20 min at 4° C, and the leukocyte layer (buffy coat) collected. Genomic DNA was isolated from the buffy coat using spin columns (Qiagen, Valencia, CA) and subjected to polymerase chain reaction (PCR)- based genotyping. The modified amplification resistant mutation system (ARMS) method originally described by Stafforini and colleagues (19) was used; primer A (5'-CTATAAATTTATATCATGCTT-3') and primer B (5'-TCACTAAGAGTCTGAATAGC-3') were used to detect wild type allele, and primer A and primer C (5'-TCACTAAGAGTCTGAATAAA-3') to detect V279F mutant allele. Annealing temperatures for PCR were 58° C during the first 10 cycles and 52° C during the next 30 cycles. For some samples, the whole length of plasma PAF-AH gene exon 9 was amplified using primer A and primer D (5'-TTTACTATTCTCTTGCTTTAC-3'), then sequenced directly with dideoxy chain termination method using primer E (5'-ATTTATATCATGCTTTTTCAAATAG-3').

Plasma PAF-AH Activity Assay

Plasma PAF-AH activity was determined based on the method of Stafforini and colleagues (21). Plasma (25 µl) was mixed with 0.475 ml of [³H]PAF (hexadecyl-2-[³H]acetyl-sn-glyceryl-3-phosphorylcholine; NEN, Boston, MA) in Tris-HCl (50 mM, pH 7.4) containing 2.0 mg/ml bovine serum albumin (Sigma, St. Louis, MO), incubated at 37° C for 10 min, and then an equal volume of 14% ice-cold trichloroacetic acid (Sigma) added to stop the reaction. After centrifugation at 1,500 × g for 10 min at 4° C, radioactivity in the supernatant was counted with a scintillation counter (model LS9800; Beckman, Fullerton, CA). The lower limit of the assay was 0.1 nmol/min/ml and the coefficient of variation 0.27.

Statistical Analysis

Plasma PAF-AH activity was presented as mean ± SD. Parameters between groups were compared using either parametric (Student's t test, one-way analysis of variance), nonparametric analysis (Mann-Whitney U test, Kruskal-Wallis test), or both. Allele frequency and genotype distribution in different groups were compared using ² analysis. p Values less than 0.05 were considered significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Patient Features

According to age, gender, age at onset, serum IgE levels, serum cholesterol levels, and dose of inhaled or oral corticosteroids required to control asthmatic symptoms, nonatopic asthmatics were significantly older (58 ± 13 yr) compared with atopic (42 ± 17 yr) or healthy subjects (40 ± 11 yr), and had later-onset asthma, lower serum IgE, and higher serum cholesterol levels (Table 1). Gender ratio differed between atopic and nonatopic asthmatics, but we found no gender-oriented difference of V279F allele frequency (male 20.4% versus female 21.2%) or of plasma PAF-AH activity (male 34.4 ± 23.1 versus female 34.3 ± 19.9 nmol/min/ml) in 111 control subjects.

PAF-AH Genotypes

The accuracy of ARMS method was confirmed by direct nucleotide sequencing of exon 9 in random samples (Figure 1). To increase the power of analysis, we genotyped an additional 106 healthy subjects (52 male, 54 female). The distribution of V279F genotypes agreed well with that predicted based on Hardy-Weinberg equilibrium both in patients with asthma and in control subjects. The allele frequency of V279F mutation was 21.7% in control subjects (n = 217) and 18.6% in asthmatics (n = 279), showing no statistically significant difference (² = 1.36, df = 1, p = 0.26) (Table 2). We also found no difference of allele frequency between atopic and nonatopic asthmatics (² = 2.53, df = 1, p = 0.13), or among patients with mild, moderate, and severe asthma (² = 2.32, df = 2, p = 0.26). Even when control subjects and nonatopic asthmatics were combined and compared with atopic asthmatics, the difference did not reach statistical significance. V279F homozygotes were found in 12 of 217 healthy subjects and 12 of 279 asthmatics (three atopic and nine nonatopic, four mild, seven moderate, and one severe).

View larger version (99K): [in this window] [in a new window]   View larger version (25K): [in this window] [in a new window]   Figure 1.   (A) Representative results of plasma PAF-AH genotyping by ARMS method. Lane 1: 100 bp ladder. Lanes 2, 4, 6: amplified fragments of wild type alleles. Lanes 3, 5, 7; amplified fragments of V279F mutant alleles. (B) Nucleotide sequences of wild type homozygote (left), V279F heterozygote (center), and V279F homozygote (right).

Plasma PAF-AH Activity

Among 111 normal subjects, plasma PAF-AH activity was absent (0.0 ± 0.0 nmol/min/ml) in V279F homozygotes (n = 4); mean PAF-AH activity in V279F heterozygotes (22.4 ± 15.0 nmol/min/ml, n = 38) was about half of that in wild type homozygotes (42.9 ± 19.8 nmol/min/ml, n = 69) (Figure 2). A similar relationship between genotype and enzyme activity was seen in the asthma group. We found, however, very low PAF-AH activity (< 2 nmol/min/ml) in five subjects with at least one wild type allele. The whole exon 9 nucleotide sequence was determined in these cases, and we confirmed that they are either heterozygotes (four cases) or wild type homozygote (one case) of V279F allele without additional mutation in exon 9. 

View larger version (11K): [in this window] [in a new window]   Figure 2.   Plasma PAF-AH genotypes and enzyme activity in healthy subjects. Mean ± SD. *p < 0.01, compared with wild type. #p < 0.01, compared with V279F heterozygote.

Cases were stratified by plasma PAF-AH genotype, and enzyme activity in plasma was compared for control subjects, atopic and nonatopic asthmatics, or among patients with different asthma severity (Figure 3). We found no significant difference due to asthma presence, type, or severity in either wild type homozygotes or V279F homozygotes.

View larger version (17K): [in this window] [in a new window]   Figure 3.   Analysis of plasma PAF-AH activity in control and asthmatic subjects followed by stratification by V279F genotype. PAF-AH activity was compared in (A) wild type homozygotes or (B) V279F heterozygotes. No V279F homozygotes had plasma PAF-AH activity. Data are means ± SD.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Because PAF activates platelets, neutrophils, eosinophils, macrophages, and vascular endothelial cells at very low concentration, the clearance of PAF such as plasma PAF-AH should be an important anti-inflammatory mechanism in terminating allergic and nonallergic inflammation. Recombinant plasma PAF-AH reportedly suppresses PAF-induced inflammation in mice (10). Our study, however, demonstrated that naturally occurring genetic deficiency of plasma PAF-AH and asthmatic phenotypes do not cosegregate in the Japanese population.

Physiological and pathological factors affect plasma PAF-AH levels; smoking (12), cerebral/coronary vascular disease (11, 17), and rheumatoid arthritis (15) increase plasma PAF-AH activity, whereas sepsis (16), systemic lupus erythematosus (13), estrogen (14), and other factors decrease it. One of the most important factors in determining plasma PAF-AH activity is serum lipoproteins as they are the PAF-AH carrier in plasma (22). In the Japanese population, the genetic factor is apparently an independent determinant of plasma PAF-AH activity. Stafforini and colleagues (19) reported that the allele frequency of V279F mutation was 17.5% in 127 random Japanese subjects, which agreed well with our result (21.7% in 217 subjects). Plasma PAF-AH activity correlated inversely with the number of mutated alleles. Several samples in our study, however, demonstrated quite low enzyme activity despite the presence of wild type alleles. We searched for another function-eliminating mutation in exon 9 because the exon encodes catalytic center Ser 272 and nearby amino acids; a novel missense mutation Q281R was reported in this region by Yamada and Yokota (23). We did not, however, find either Q281R or any other novel mutation in exon 9 of these samples. It remains unclear if an unknown mutation exists in other exons or if nongenetic factors diminish enzyme activity in these cases.

It is unclear whether deficiency or low plasma PAF-AH is associated with asthma. Miwa and coworkers (18) reported that mean plasma PAF-AH activity in atopic asthmatic children with mild symptoms (1.40 ± 0.49 nmol/min/50 µl) or moderate symptoms (1.21 ± 0.53 nmol/min/50 µl) did not differ from that in healthy children (1.25 ± 0.49 nmol/min/50 µl), whereas children with severe asthmatic symptoms exhibited low enzyme activity (0.99 ± 0.59 nmol/min/50 µl serum). They also found high prevalence of PAF-AH deficiency (16.7%; 3 of 18 cases) compared with healthy children (3.8%; 8 of 211 cases). In contrast, Tsukioka and coworkers (24) observed significantly lower plasma PAF-AH activity in adults with asthma (12.7 ± 8.3 nmol/min/ml) than in normal subjects (21.2 ± 6.8 nmol/min/ml), and asthma severity was not associated with enzyme activity.

The sample size in our study (279 cases, 111 controls) was equivalent to that in Miwa's (175 cases, 211 controls) or Tsukioka's (137 cases, 106 controls). Our study included enough subjects to detect a 30% change of PAF-AH activity with the power of 95%. We also had the advantage of V279F mutation genotyping which enabled us to identify heterozygotes and determine allele frequency. Recently, Hiramoto and colleagues (25) found that V279F mutation prevalence in stroke patients (43.4%) was higher than in the normal Japanese population (25.4%); the difference was mostly due to the difference of heterozygote frequency (39.2% versus 22.4%), but not homozygote frequency (4.2% versus 3.0%). We did not find any deviation of allele or genotype distribution in the population we studied, indicating that the plasma PAF-AH gene is not a major gene for asthma and its phenotypes in Japanese adults. It is possible that we missed small differences of allele frequency between asthmatics and the normal population. The sample size in our study has the power ( = 0.8) to detect a 10% difference of allele frequency between groups, but we would need approximately 1,000 subjects for each group to detect a 5% difference of allele frequency, although the biological and clinical significance of such a small difference would be ambiguous.

By genotyping, we could evaluate the effects of nongenetic factors on plasma PAF-AH activity, which, after stratification by the genotypes, demonstrated no significant difference of PAF-AH activity between healthy control subjects and either atopic or nonatopic asthmatics, or among patients with different disease severity. We thus believe that asthmatic inflammation has little or no effect on plasma PAF-AH levels. This is also supported by the fact that PAF-AH activity was consistent during and after acute exacerbation of asthma, observed both by Miwa and Tsukioka (18, 24). Interestingly, Triggiani and coworkers (26) found that PAF-AH activity in bronchoalveolar lavage fluid was significantly lower in asthmatic patients than in normal subjects or in patients with interstitial lung diseases. Their study, done on a Caucasian population, was unrelated to V279F mutation, and thus suggests the presence of nongenetic factors that decrease PAF-AH activity in the airway but not in plasma.

Our results indicate that plasma PAF-AH is not as biologically essential to asthma pathophysiology as has been believed, although PAF receptor antagonists are apparently useful in asthmatics and in animal asthma models (27, 28). It is possible that PAF is degraded by intracellular isoforms of PAF-AH. Red blood cell PAF-AH activity, for example, accounts for about 35% of enzyme activity in whole blood and remains unchanged in subjects with plasma PAF-AH deficiency. Yoshida and coworkers, however, demonstrated that whole blood containing red blood cells from subjects with plasma PAF-AH deficiency still showed prolonged PAF half-lives (29). There thus may be other unknown extracellular isoforms of PAF-AH in the airway; Triggiani and coworkers, based on different susceptibility to a panel of serine protease inhibitors, concluded that PAF-AH activity in bronchoalveolar lavage fluid was derived from an enzyme other than plasma PAF-AH (26). Further study is required for PAF metabolism and clearance in the airway.

Our results demonstrated that V279F mutation itself is not the predisposing factor of atopy, asthma, or asthma severity. It is still possible that PAF-AH deficiency determines another asthma phenotype. It will be especially interesting and clinically important to study if PAF receptor antagonists are more effective in PAF-AH-deficient patients with asthma.