Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Over the past decade, evidence has accumulated pointing to cysteinyl-leukotrienes (cys-LTs) (1) as the key mediators in bronchial asthma (4). They are potent bronchoconstrictors, induce mucus hypersecretion and airway edema, and recruit eosinophils into human lungs. Their engagement in the pathomechanisms of the disease has been confirmed recently by anti-leukotriene drugs effective in asthma (5).

Cys-LTs are generated upon cellular activation, when arachidonic acid, a 20-carbon polyunsaturated fatty acid, is released from nuclear membrane phospholipids by the action of cytosolic enzyme phopholipase A₂ and is converted into an unstable intermediate, leukotriene (LT) A₄, by 5-lipoxygenase in the presence of its activating protein, FLAP. The integral perinuclear membrane protein LTC₄ synthase (LTC₄S) conjugates LTA₄ with reduced glutathione to form the intracellular product LTC₄. After carrier- mediated cellular transport, sequential cleavage of glutamic acid and glycine provides the final derivatives LTD₄ and LTE₄. LTC₄S was cloned, and its gene was localized to the distal part of the long arm of chromosome 5 (5q35) (8, 9), the region telomeric to the loci of cytokines and receptors implicated in allergy and bronchial hyperresponsiveness. The mechanism by which the enzyme activity is regulated remains unclear. The enzyme is under transcriptional control, which limits its expression to eosinophils, basophils, mast cells, macrophages, platelets, and endothelial cells (10).

Aspirin-intolerant asthma (AIA), a distinct clinical syndrome (11) that affects about 10% of adult asthmatics, appears to be unusually dependent on cys-LT overproduction. At baseline, patients with AIA excrete high amounts of cys-LT in urine (14, 15). After ingestion of aspirin and other cyclooxygenase inhibitors (16), they release cys-LTs into lungs (17) and further increase the excretion of urinary LTE₄ (13). Bronchial biopsy studies revealed a marked overexpression of LTC₄S in the AIA lung compared with aspirin-tolerant and normal lung (18). The overexpression of LTC₄S in the bioptates was accompanied by elevated cys-LT levels in bronchoalveolar lavage fluid of patients with AIA and was the only biochemical or cellular parameter measured in AIA that correlated with bronchial hyperresponsiveness to inhaled lysine-aspirin (19).

Recently, we described a genetic polymorphism of the 5' untranslated region of LTC₄S (20). The single nucleotide polymorphism (SNP) consists of two common alleles corresponding to an A to C transversion at the site 444 nucleotides upstream of the translation start. In a pilot clinical study, we found the ₄₄₄C allele frequency more common in aspirin-sensitive asthmatics than in patients with aspirin-tolerant asthma (ATA) or normal subjects. We wondered whether the allelic variant creates a new binding site for transcription factors, whether it affects the rate of transcription in vitro, and if so, whether this is reflected in a clinical setting. The results obtained indicate that: (1) the ₄₄₄C has a new protein binding motif for the histone H4 transcription factor, overlapping its activator protein (AP)-2 consensus sequence; (2) transfection of cells with ₄₄₄C promoter construct augments expression of reporter genes; (3) the ₄₄₄C allele frequency is increased in AIA, especially in patients with a severe steroid-dependent type of disease; (4) patients who suffer from AIA have enhanced LTC₄S messenger RNA (mRNA) expression in blood eosinophils; and (5) patients with AIA, carriers of ₄₄₄C allele, respond to aspirin challenge with increased LTE₄ urinary excretion.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Gel Shift Assays

Experiments were performed using a gel shift assay core system (Promega, Madison, WI), according to the recommendation of the manufacturer. Double-stranded oligonucleotides₄₄₄A-LTC₄S allele, GGA TGG GGA CAG GGA ACA GAT, ₄₄₄C-LTC₄S allele, GGA TGG GGA CCG GGA ACA GAT, and H4TF-2 consensus, CGG ACC GGG GGA GAA CCwere synthesized. The others, AP-2 consensus, GAT CGA ACT GAC CGC CCG CGG CCC GT and SP1 consensus, ATT CGA TCG GGG CGG GGC GAG, were included in the kit. The oligonucleotides were end-labeled with ³²P using T4 polynucleotide kinase reaction. The nuclear protein extract was prepared (21) from peripheral blood eosinophils isolated by magnetic immunoselection. The cells (2 million) were incubated with hypotonic buffer and subsequently lysed with 1% NP40, and the nuclear pellet was recovered by centrifugation. After resuspension in a high salt buffer and separation in a centrifuge, the supernatant containing nuclear proteins was stored at 70°C.

Oligonucleotide Decoy Experiment

Promyelocytic HL-60 cell line was used. Expression of LTC₄S in HL-60 cells, differentiated into a granulocyte pathway by 5-d incubation in 1.3% dimethyl sulfoxide (LTC₄S genotype AC), was confirmed by reverse transcriptase/polymerase chain reaction (RT-PCR) for LTC₄S and by accumulation of cys-LTs in supernatants. LTC₄S mRNA was also detectable by RT-PCR in undifferentiated HL-60 cells but in a lower concentration (data not shown). Cells (7.5 × 10⁶/point) were transfected with use of lipofectamine plus (GIBCO BRL, Rockville, MD) at a concentration of 40 µl/25 ml of RPMI medium with 10% fetal calf serum and 100 nM of double-stranded H4TF-2 oligonucleotide. The sequence of the decoy oligonucleotide contained both H4TF-2 and AP-2 responsive elements (Figure 1). Controls were mock transfected. After 24 h incubation, a fresh medium (2 ml) was replaced with addition of 50 µM arachidonic acid, and the cells were stimulated with 5 µM FMLP for 30 min. Cys-LTs were measured in supernatants using an enzyme-linked immunosorbent assay kit (Bühlmann Laboratories, Switzerland).

View larger version (14K): [in this window] [in a new window]   Figure 1.   Sequences of oligonucleotides used for gel shift assays. An SNP of an LTC4S fragment (bases, 457 to 434 from the gene translation start) is underlined in the 444A-LTC4S and 444C-LTC4S oligonucleotides. For the AP-2 oligonucleotide, a complementary string is aligned. The same H4TF-2 oligonucleotide was used for the decoy experiment. The gray rectangle marks the H4TF-2 consensus core sequence. The border delineates potential AP-2 transcription factor binding sites.

Construction of Expression Vectors

For amplification of a promoter fragment subsequently cloned into expression vectors, primers 5'-TCC ATT CTG AAG GCA AAG GC and 5'-GGA GAC CGC CTC ACC ACT T were used. The amplified 297-bp fragment was cloned into a PCR 2.1 TA cloning vector (Invitrogene, Carlsbad, CA). For construction of a -galactosidase expression vector, a PCDNA3.1MycHisLacZ() plasmid with reverse polylinker sequence (Invitrogene) was cut with restriction endonucleases NruI and SnaBI, creating blunt ends. Digestion removed about half of the original plasmid cytomegalovirus (CMV)-derived promoter sequence. The vector was dephosphorylated with calf intestine phosphatase to prevent religation. The clones of LTC₄S alleles in PCR 2.1 vector were cut with endonucleases SpeI and EcoRV at the polylinker sites of the plasmids, 33 and +17 nucleotides distant from the LTC₄S insert. After ligation, the correct orientations of inserts were verified by PCR.

Cell Transfection

COS-7 cells (American Type Culture Collection, Rockville, MD) were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal calf serum. Cells were seeded in 60-mm tissue culture dishes at the density of 5 × 10⁴/plate, 2 d before experiments. Diethylaminoethyl (DEAE)-dextran was used for transfection of the cells (22). Cells were incubated with DEAE-dextran-plasmid precipitate at the concentration of 200 µg/ml and 5 µg of each for 2 h. To detect the -galactosidase activity, cells were washed twice in phosphate-buffered saline (PBS), fixed for 5 min at 4°C in 3.7% formaldehyde in PBS, and stained for 2 h in X-gal staining solution (80 mM Na₂HPO₄, 20 mM NaH₂PO₄, 1.3 mM MgCl₂, 3 mM K₃Fe[CN]₆, 3 mM K₄Fe[CN]₆, 1 mg/ml X-gal). The method of computer-assisted image analysis was used to evaluate the activity of -galactosidase in transfected cells. The images were collected with a color CCD camera (DICD; World Precision Instruments, Sarasota, FL) mounted on an inverted microscope and attached to a personal computer. The data were then analyzed by the SigmaScan Pro 2.0 image measurement system (Jandel Scientific, San Rafael, CA). At least 100 positively transfected cells were counted for each experiment.

Subjects Studied

The study was performed in 76 unselected asthmatic patients sensitive to aspirin (AIA; average age, 45 yr; 19 men and 57 women), 110 unselected asthmatics who tolerated aspirin well (ATA; average age, 37 yr; 44 men and 66 women), and 75 health control volunteers (average age, 35 yr; 29 men and 46 women). Within the 9 mo preceding the study, those with AIA had undergone the inhaled provocation test with L-lysine aspirin, which clearly documented their sensitivity to aspirin and established the dose that produced a 20% fall in forced expiratory volume in one second (PD₂₀) (16). ATA subjects and healthy controls all took aspirin or other nonsteroidal anti-inflammatory drugs during the previous year without any adverse symptoms. The two patient groups were similar in age and sex, with no significant differences in the duration of asthma, spirometry, serum immunoglobulin E, or duration and dosage of corticosteroid medication.

Healthy subjects had no history of chronic respiratory disease and were taking no medication.

Genotyping of the Subjects Studied

Genomic DNA samples were obtained from the peripheral blood of 76 AIA, 110 ATA, and 75 healthy individuals. The promoter region of LTC₄S was amplified with the same primer pair, as used for the expression vector cloning experiments. After digestion with MspI restriction endonuclease, the products were typed on acrylamide gels by electrophoresis as described previously (20).

Peripheral Blood Eosinophil Studies

For eosinophil isolation, blood was collected on ethylenediaminetetraacetic acid from 45 patients with AIA, 63 patients with ATA, and 16 healthy volunteers. Asthmatics were not receiving oral corticosteroid medication. Purification of eosinophils by density gradient centrifugation and CD16 immunomagnetic negative selection were performed as previously described (23). Anti-CD16 monoclonal antibody bound to micromagnetic beads and a magnet-activated cell sorter system were obtained from Miltenyi Biotech (Auburn, CA). Diluted blood was layered onto 1,119 and 1,077 g/ml histopaque gradient (Sigma Chemical Co., St. Louis, MO), and centrifuged at 300 × g for 30 min at 20°C. The upper layer containing the mononuclear cells was discarded, and the lower layer, above the red blood cells, was collected. Remaining red blood cells were removed by lysis with 155 mM NH₄Cl and 10 mM KHCO₃ buffer. Eosinophil purity was > 90% (in May-Grünwald-Giemsa stained preparation) and viability was > 90%. Freshly separated eosinophils were incubated in DMEM at a cell density of 1 × 10⁵ cells/200 µl in flat-bottom, polystyrene culture wells for 18 h at 37°C under 5% CO₂ with or without 20 ng/ml of human recombinant interleukin (IL)-5 (Genzyme, Cambridge, MA). After incubation, total cell RNA was isolated using standard procedure (TRI reagent; Sigma Chemical Co.). Most of the experiments were done in duplicates. Total cell RNA was reverse transcribed using a LTC₄S gene specific primer, and a competitive PCR was performed with addition of a constant amount of genomic DNA (3 ng  900 copies) as an internal control. The primers used for simultaneous amplification of complementary DNA (cDNA) and genomic fragments of LTC₄S were 5'-CCC GAG TTC GAG CGC GTC TA and 5'-GCC CTG GAA GTA GCG GAG GC. The reaction products were 159-bp cDNA and 345-bp genomic DNA fragments. Quantitative results were obtained using the automated acrylamide gel electrophoresis of products with laser fluorometry (ALF-Express; Pharmacia, Uppsala, Sweden), and the sensitivity threshold was estimated by a serial dilution to 50 molecules per sample. The number of cDNA copies in an aliquot of reverse transcribed mRNA, corresponding to 0.1 µg of total RNA, was calculated from the ratio of areas of the peaks corresponding to cDNA and genomic DNA.

LTE₄ Excretion in Urine

LTE₄ was measured in unpurified samples (15) by direct enzyme immunoassay (Cayman Chemical, Ann Arbor, MI) and expressed in picograms per milligrams of creatinine in a group of 24 AIA patients at baseline, just before inhalation of a single PD₂₀ dose of lysine-aspirin (17), and every 2 h for 6 h after inhalation.

Statistical Analysis

To determine the relative risk of AIA associated with the ₄₄₄C LTC₄S allele, its 5% and 95% confidence intervals were calculated according to Sheehe (24).

Median LTC₄S transcript abundances per 0.1 µg of total cell RNA were reported with a 25th to 75th percentile interval, within group comparisons were tested using Wilcoxon's matched pairs test, and between group comparisons were tested using Mann-Whitney U test.

Average LTE₄ excretions in urine per milligram of creatinine were reported with standard deviations, and the same statistical tests were used for comparisons as for transcript abundances.

The protocol of the study was approved by the Jagiellonian University Ethical Committee, and informed consent was obtained from each patient.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Gel Shift Assays

Incubation of HeLa nuclear extracts with the 21-bp fragments of both alleles of LTC₄S promotor led to the formation of protein-DNA complexes. When resolved on the acrylamide gel, one of them corresponded to the control complex of recombinant AP-2 protein and tested LTC₄S oligonucleotide fragment. The specificity of the protein-DNA interaction was assessed by competition studies with an unlabeled AP-2 consensus sequence oligonucleotide. Competition between the LTC₄S oligonucleotide and the AP-2 oligonucleotide resulted in about 3-fold less binding of LTC₄S allele ₄₄₄A, but only a 20% weaker signal for the ₄₄₄C allele. This difference was, however, negligible in experiments in which recombinant AP-2 protein was used instead of HeLa nuclear extract. The AP-2 consensus oligonucleotide excess attenuated the LTC₄S oligonucleotide- protein complex signal to the same extent for either of the allelic variants.

The study of nuclear protein interactions of DNA-eosinophils revealed further differences between allelic variants of LTC₄S promoter (Figure 2). H4TF-2 oligonucleotide (1) showed strong binding with eosinophil nuclear extract and (2) attenuated binding of LTC₄S ₄₄₄C oligonucleotide with nuclear extract proteins. This interaction was even greater when nuclear protein complexes with H4TF-2 oligonucleotide were treated with LTC₄S ₄₄₄C oligonucleotide. This result confirmed that a difference between common alleles of the LTC₄S promoter depended on presence of an additional binding site for transcription factor, created by transversion in the ₄₄₄C allele. The changed sequence corresponded to the responsive element of histone H4 promoter.

View larger version (90K): [in this window] [in a new window]   Figure 2.   Gel shift assay of DNA-eosinophil nuclear protein interactions. The aliquots of eosinophil nuclear extracts (2 µg each) were incubated with 10 pmols of 5' end-labeled [32P], double-stranded 444A-LTC4S, 444C-LTC4S, and H4TF-2 oligonucleotides for 20 min on ice. A twentyfold excess of unlabeled oligonucleotides was added to some incubates, as indicated in the table. After 20 min of further incubation at room temperature, the samples were loaded on modified acrylamide gel (HydroLink; Bioprobe, Montreuil, France) and separated by electrophoresis for 3 h at 300 V in 0.5× TBE buffer. The gel was fixed after the electrophoresis, vacuum-dried, and exposed overnight to BioMax autoradiography film (Amersham, Buckinghamshire, UK). The arrow indicates competition between H4TF-2 and 444C-LTC4S for nuclear protein binding sites.

Cys-LT Production after Decoy Oligonucleotide Transfection

In six consecutive experiments, nondifferentiated HL-60 cells accumulated cys-LTs to the concentration of 112 ± 59 pg/ml per 1 million cells. The cells transfected with the double-stranded H4TF-2 oligonucleotide did not differ in cys-LT accumulation (147 ± 79 pg/ml). Differentiated HL-60 cells released severalfold more cys-LTs (974 ± 606 pg/ml), and decoy oligonucleotide inhibited biosynthesis to 68% (661 ± 524 pg/ml; P < 0.03, Wilcoxon's matched pairs test).

Transient Expression of LTC₄S  Promoter-Driven Expression Vectors

LTC₄S promoter has numerous transactivation sites, including AP-1, SP1, nuclear factor B, and cyclic AMP response element. For the in vitro expression studies, a 296-bp fragment of either allele was introduced into plasmids carrying reporter genes (25).

In four separate experiments in which COS-7 cells were transfected with the LTC₄S-CMV hybrid promoter and -galactosidase reporter gene construct, consistently higher expression of -galactosidase was produced by ₄₄₄C allele. The net difference, measured as optical density of X-gal-stained cells, for vectors carrying A and C alleles varied between transfection experiments from 3 to 57%. The mean difference of more than 100 ascertained cells was highly significant (P < 0.01) for each of the transfections.

Alellic Frequencies of the LTC₄S Gene

Studies of the region of the LTC₄S promoter revealed a diallelic polymorphism as described previously, resulting in three different genotypes (Table 1). The ₄₄₄C allelic variant was less frequent in all examined groups. There were no differences between aspirin-tolerant asthmatics (q = 0.27, n = 110) and control subjects (q = 0.27, n = 75) in the frequency of LTC₄S alleles. Patients with AIA had increased frequency of the ₄₄₄C allele (q = 0.39, n = 76). The allelic frequencies in patients with AIA differed significantly from those of healthy control subjects (² = 4.99, P = 0.025) and from those of aspirin-tolerant asthmatics (² = 6.62, P = 0.01). The relative risk of aspirin-induced asthma related to the ₄₄₄C LTC₄S allele status was 2.62 (95% confidence interval, 1.38 to 4.98), and the difference in frequency of carrier status between patients with AIA and control subjects was significant (² = 8.71, P < 0.004).

Frequency of the LTC₄S allele ₄₄₄C calculated for a subpopulation of asthmatics with mild disease not requiring systemic corticosteroids was 0.31 in patients with AIA (n = 26) and 0.33 in patients with ATA (n = 33).

Expression of LTC₄S mRNA in Peripheral Blood Eosinophils

Experiments in which reverse transcribed mRNA for LTC₄S was coamplified with 900 copies of genomic DNA, gave clear signals of cDNA in most samples from asthmatics, using a competitive PCR. Quantification of the cDNA band by calculation of signal ratio to the genomic internal control allowed us to estimate the number of copies of mRNA (Table 2). The LTC₄S mRNA message was increased in asthmatic samples when compared with control samples (control versus ATA, P < 0.01; control versus AIA, P < 0.001). Within the patients, the AIA group had a significantly higher expression (P < 0.001) of LTC₄S than did the ATA group. These differences were leveled off by incubation of peripheral blood eosinophils with IL-5. After IL-5 incubation only ATA patients had more LTC₄S transcripts (P < 0.03) than did control subjects. Incubation with IL-5 significantly (P < 0.01) increased LTC₄S mRNA in healthy control subjects and ATA patients but not in AIA patients.

Asthmatics carrying the ₄₄₄C allele had similar LTC₄S mRNAs in peripheral blood eosinophils as compared with ₄₄₄A homozygotes at base (median = 72 [24 to 271] versus 77 [23 to 243]) or after IL-5 incubation (median = 80 [44 to 207] versus 94 [39 to 298]).

Urinary Excretion of LTE₄

In AIA patients, there was no difference at base in urinary LTE₄ excretion between carriers of the allele ₄₄₄C and ₄₄₄A homozygotes. However, after provocation with inhaled lysine-aspirin at a PD₂₀ dose, increase in urinary output of LTE₄ was more intense in ₄₄₄C allele carriers (Figure 3) than in ₄₄₄A homozygotes. This difference was significant in the urine samples collected between the second and fourth hours after the challenge (1,311 ± 1,113 versus 2,144 = 1,632, P < 0.006). There was no difference in PD₂₀ of aspirin between carriers of the ₄₄₄C allele and ₄₄₄A homozygotes of LTC₄S.

View larger version (26K): [in this window] [in a new window]   Figure 3.   Urinary excretion of LTE4 in pg/mg creatinine of patients with AIA challenged with inhalatory aspirin PD20. Urine samples were collected in 2-h intervals, before the challenge and for the next 6 h. Patients were grouped according to their LTC4S genotypes; white columns represent average LTE4 and standard error of the mean for 444A-LTC4S homozygotes; black columns denote carriers of 444C-LTC4S allele.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Analysis of the promoter region of the LTC₄S gene (GeneQuest; DNA Star Co., Madison, WI) suggested that the SNP could affect binding of transcription factors to the alleles. Adenine to cytosine nucleotide transversion at the AP-2 responsive element created an additional site for the transcriptional signal of a histone H4 transcription factor-2 (26) (Figure 1).

Protein-DNA interaction studies with LTC₄S promoter fragments were in agreement with the computer prediction. We demonstrated binding of the AP-2 transcription factor to a DNA fragment corresponding to the nucleotides 457 to 434 of the LTC₄S promoter. This interaction occurred with either allelic variant, though binding of nuclear extracts was stronger with the allele ₄₄₄C. An oligonucleotide having H4TF-2 consensus sequence interfered with binding of the AP-2 transcription factor to the allele ₄₄₄C fragment but not to the allele ₄₄₄A. In the ₄₄₄C allele, the presence of an additional protein binding motif, that of histone H4 transcription factor-2, was confirmed by strong interference between labeled H4TF-2 and a ₄₄₄C fragment when tested in nuclear extracts of eosinophils. This combination of the two overlapping protein binding sites had properties not only detectable by gel shift assay but also measurable by LTC₄S activity. In the experimental blockade of transcription factors by a specific oligonucleotide decoy (27), we confirmed that the promoter region encompassing the SNP functioned as a transcription enhancer. Addition of an excess of the synthetic double-stranded oligonucleotide to the LTC₄S positive cells bound nuclear transcription factors and diminished LTC₄S expression.

The AP-2 transcription factor has rather weak DNA binding properties and acts as a coactivator of gene expression. It mediates cell differentiation in response to retinoic acid and in concert with other DNA binding proteins (28). The H4TF-2 protein is present in cells only during the DNA synthesis phase of the cell cycle (29). H4TF-2 is deregulated and frequently overexpressed in tumor cells. Promoters with H4TF-2 responsive elements have been found in relatively few genes (30). Among them is the eosinophil peroxidase (EPO) gene, whose expression is limited to eosinophils (31). Interestingly, EPO participates in the mechanism of bronchial hyperresponsiveness by inhibition of muscarinic receptors M2 in human airways (32).

Experiments in which fragments of the LTC₄S promoter encompassing the SNP were linked to the reporter gene and transiently expressed in cell cultures showed rather weak transactivating properties of the construct. We failed to identify the functional component of the promoter of the enhancer type. Vectors containing the LTC₄S promoter as the sole sequence attached to the green fluorescent protein reporter gene had very low expression when tested on COS-7 and human melanoma B16 cell lines, in contrast to the positive control having a CMV promoter (data not shown).

The alleles of LTC₄S, nevertheless, consistently differed in expression when linked to the reporter gene. By hybrid promoter studies in which we increased a reporter gene signal through coupling of a LTC₄S fragment to the CMV promoter fragment, we managed to register the difference in their transcriptional activities. The ₄₄₄C allele of LTC₄S led to higher expression of -galactosidase in transfected COS-7. On the basis of cell culture studies, the effect of the ₄₄₄C allele could be quantified on average as a 25 to 30% increase in transcription rate of the gene. However, the pathophysiologic consequences of the ₄₄₄C allele overexpression are more profound.

In the subjects studied by us, the ₄₄₄C allele was associated with AIA but not with ATA or a healthy state. However, allelic association seemed to depend on the disease severity (33). When less frequent, patients with mild AIA not requiring chronic oral corticosteroids medication were compared with matched patients with ATA and no differences in frequency of ₄₄₄C allele could be demonstrated, despite enhanced transcription of the gene in circulating eosinophils of AIA positive for ₄₄₄C.

Studies on LTC₄S mRNA in peripheral blood eosinophils revealed that asthmatics, especially patients with AIA, have an increased number of the gene transcripts. The ₄₄₄ C allele, by interaction with a specific transcription factor, had an additive effect on the transcription, generally upregulated, and significantly higher in the AIA group than in the two other groups studied. The effect of ₄₄₄C allele on LTC₄S expression was minor when compared with the induction of the gene by an unknown triggering factor. This observation is in agreement with nonobligatory ₄₄₄C carrier status, as a predisposing factor to AIA. Induction of LTC₄S expression by IL-5 did not depend on genotype but rather on basal transcript expression. Thus in AIA, the LTC₄S gene is upregulated, probably due to circulating inflammatory mediators, and the promoter genotype has only a permissive, modulating effect. Eosinophilia is a prominent feature in blood, airways, and nasal polyps of aspirin-sensitive patients (12, 34). In asthmatics, bronchi eosinophils are the main source of LTC₄S (18, 19); the enzyme is also present in alveolar macrophages (35). The enzyme is markedly overexpressed in the bronchi of patients with AIA, perhaps because eosinophils are primed by various cytokines in situ (19). In addition, as shown here, the transcription is already upregulated in eosinophils circulating in blood, before they penetrate to the airways. During the development of eosinophils from their progenitors, the acquisition of LTC₄S occurs together with eosinophil granule development as a maturation-related protein (36). Our results suggest existence of a mechanism by which a transcription factor binding to a histone H4 promoter motif, operating within the progenitor cells of dividing eosinophils, enhances LTC₄S transcription in carriers of a common ₄₄₄C allele of this gene. This does not change overall cellular distribution of the LTC₄S as its transcription is effectively limited to mononuclear bone marrow-derived eosinophils, basophils, and platelets. Overexpression of the LTC₄S gene could develop in response to a hemopoietic signal sent from inflamed bronchi to the progenitors of bone marrow (37). These hypotheses remain to be tested.

Enhanced urinary basal LTE₄ excretion, previously reported (13) in AIA, was unrelated to LTC₄S polymorphism. However, provocation with aspirin led to a significant increase in urinary output of LTE₄, but only in carriers of the ₄₄₄C allele. These results are consistent with the hypothesis that aspirin turns on the cys-LT pathway, whereas genetic overexpression of the key enzyme is enhanced in carriers of the allelic variant ₄₄₄C. Determination of the LTC₄S genetic polymorphism seems, therefore, warranted in future studies on identification of those patients who can expect the most benefit from anti-leukotriene drugs.

In conclusion, the phenotype of AIA is associated with an allelic variant, ₄₄₄C, of LTC₄S. This allelic variant has an additional responsive element to histone H4 transcription factor-2, which increases the transcription rate of the gene both in vitro and in vivo. These findings explain certain features of cys-LT overproduction, including a hallmark of AIA. Our results also suggest existence of an additional unknown triggering factor, like persistent viral infection, that induces the gene and is largely responsible for the enhanced transcription by blood eosinophils and augmented basal excretion of urinary LTE₄.