intranasal salmeterol inhibits allergen-induced vascular permeability but not mast cell activation...

Intranasal salmeterol inhibits allergen-induced vascularpermeability but not mast cell activation or cellular infiltration

D. PROUD, C. J. REYNOLDS, L. M. LICHTENSTEIN, A. KAGEY-SOBOTKAand A. TOGIAS

Division of Clinical Immunology, Department of Medicine, Johns Hopkins University School of Medicine, Baltimore,Maryland, USA

Summary

Background Salmeterol is a long-actingb2-adrenergic agonist that is widely used in thetreatment of asthma. It has been suggested that non-bronchodilator actions of salmeterolmay contribute to its efficacy.Objective To further evaluate the potential non-bronchodilator actions of salmeterolinvivo, using a model of nasal challenge with allergen.Methods Twelve asymptomatic subjects with seasonal allergic rhinitis participated in arandomized, double-blind, placebo-controlled crossover trial of the effects of a single doseof 100mg of salmeterol on the response to allergen challenge. Sneezing and symptomscores, and levels of histamine and albumin in nasal lavages, were measured throughout theprotocol. Concentrations of tryptase, prostaglandin D2 and lysozyme were measured duringthe acute allergic response, while levels of IL-3, IL-5 and IL-8 were measured at later timepoints. Numbers of eosinophils and of total white blood cells were also recorded.Results Salmeterol did not affect sneezing or symptom scores at any point. Duringthe immediate response to allergen challenge, mast cell activation, reflected by concen-trations of histamine, tryptase and prostaglandin D2, and serous glandular secretion,assessed by measurements of lysozyme, were unaffected by salmeterol treatment butvascular permeability, reflected by concentrations of albumin in nasal lavages, wassignificantly reduced. At later time points, salmeterol had no effect on levels of histamineor albumin and did not affect cellular infiltration. Concentrations of IL-3, IL-5 and IL-8were not increased by allergen challenge in these subjects, so the effects of salmeterol couldnot be evaluated.Conclusions Treatment with a single dose of salmeterol had no effect on activation ofmast cells or cellular infiltration but inhibited vascular permeability. The ability ofsalmeterol to inhibit antigen-induced vascular permeability may contribute to its therapeuticefficacy in asthma.

Keywords: asthma, rhinitis, salmeterol,b2-adrenergic agonist, mast cells, glands, vascularpermeability, cytokines

Clinical and Experimental Allergy, Vol. 28, pp. 868–875. Submitted 27 July 1997; revised14 October 1997; accepted 20 November 1997.

Introduction

Salmeterol is a potent, topically activeb2-adrenoreceptoragonist that is widely used in the treatment of asthma [1].

Like all inhaled b2-adrenoreceptor agonists, salmeteroldirectly induces relaxation of bronchial smooth musclebut, in comparison to other available medications in thisclass, salmeterol has a slower onset of action (<10–20 min)and a significantly extended duration of action, with a singleinhaled dose of salmeterol inducing bronchodilation thatlasts at least 12 h in patients with asthma [2].

Clinical and Experimental Allergy,1998, Volume 28, pages 868–875

868 q 1998 Blackwell Science Ltd

Correspondence: Dr D. Proud, Johns Hopkins Asthma and Allergy Center,5501 Hopkins Bayview Circle, Baltimore, MD 21224–6801, USA.

In addition to its direct bronchodilator effects, salmeterol,like other b2-adrenoreceptor agonists, has been shown toinhibit IgE-mediated release of bronchospastic mediatorsfrom human lung mast cellsin vitro [3], and it has beensuggested that the ability of salmeterol to inhibit allergen-induced increases in bronchial hyperreactivity is unrelatedto its bronchodilator effects [4]. These observations,together with data demonstrating that salmeterol has anti-inflammatory properties in animal models [5], have raisedthe possibility that non-bronchodilator properties of salme-terol may contribute to the clinical efficacy of this drug inasthma. Despite this suggestion, there are no clear data tosupport that salmeterol exerts any clinically significant anti-inflammatory effects in humans.

To gain further insights into the potential non-broncho-dilator actions of salmeterol in humans, we have performeda study using a well-characterized model of nasal challengewith allergen in which indices of mast cell activation,glandular secretion, vascular permeability and cellularinfiltration can be monitored in recovered nasal lavages.We demonstrate that administration of a single dose ofsalmeterol inhibits antigen-induced increases in vascularpermeability but does not alter mast cell activation, serousglandular secretion or cellular infiltration.

Materials and methods

Subjects

Twelve asymptomatic subjects (eight female) with a historyof seasonal allergic rhinitis and a positive skin test to eithermixed grasses or ragweed were studied in the pollen-freewinter months. Subjects were challenged with the allergento which they were the most sensistive. None of the subjectshad taken glucocorticoids within 1 month or astemizolewithin the 3 months prior to study. No other medicationshad been used within 1 week of study entry. All subjectsgave informed consent prior to participating in the study andthe protocol was approved by the Institutional ReviewBoard of the Johns Hopkin Bayview Medical Center.

Protocol

The study was a randomized, double-blind, placebo-con-trolled crossover trial of the effects of a single dose of 50mgper nostril (100mg total) of salmeterol. The study protocol isoutlined in Fig. 1. The two treatment arms of the study wereseparated by 1–2 weeks. Upon entering the laboratory foreach study arm, subjects underwent five nasal lavages, eachwith 10 mL (5 mL per nostril) of prewarmed (378C) lactatedRinger’s solution (Kendall McGaw Laboratories, Irvine,CA, USA) to reduce cell-free mediator levels to a low,stable baseline. Each lavage was expelled into a collecting

basin and discarded. To reduce the mucosal oedema thatoccurs as a result of allergen challenge and can interferewith the recovery of nasal lavage fluid, subjects thenreceived 2 sprays per nostril of 0.5% oxymetazoline hydro-chloride (Rite Aid Corp., Harrisburg, PA, USA). It has beendemonstrated previously that this treatment has no effect onmediator production or sneezing in this challenge model[6,7]. Moreover, oxymetazoline has been shown to have noeffect on increases in vascular permeability induced byvasoactive mediators [8,9]. Ten minutes after administrationof oxymetazoline, an additional nasal lavage was performedand discarded. As a control, subjects then received a nasalchallenge using the diluent vehicle for the antigen extract(Greer Laboratories, Lenoir, NC, USA). Ten minutes afterthe challenge, two nasal lavages were performed in rapidsuccession and retained for processing and analysis (seebelow). Following the diluent challenge and lavages, sub-jects received intranasal placebo or salmeterol, administeredas two sprays per nostril, from freon propelled canisters thathad been placed in intranasal Beconase metered dosenasal adapters. After waiting 30 min to permit optimumreceptor occupancy by the drug, subjects were challengedwith three increasing doses [10, 100 and 1000 proteinnitrogen units (PNU)] of appropriate allergen (GreerLaboratories, Lenoir, NC, USA). Each allergen challengewas separated by 12 min with a single nasal lavage beingperformed 10 min after the first two challenge doses and twolavages being performed after the third antigen challenge.Subjects also returned to the laboratory at 4, 8, and 24 h afterantigen challenge and two nasal lavages were performed ateach time point.

Symptom scores and sneeze counts were recorded by eachsubject at the start of the protocol and 10 min after challengewith diluent and each dose of allergen. Symptom scores andsneeze counts were also recorded prior to the lavages at 4, 8,and 24 h, with sneezes recorded as the cumulative numberexperienced since the last score and count was recorded.Subjects rated symptoms on a scale of 0¼ absent to 5¼ theworst imaginable for each of the following categories: nasalcongestion, rhinorrhoea, nasal irritation and ‘other’. Thus,the maximum possible symptom score was 20.

Intranasal salmeterol 869

q 1998 Blackwell Science Ltd,Clinical and Experimental Allergy, 28, 868–875

Fig. 1. Schematic of the study protocol. Allergen doses are inprotein nitrogen units (PNU). All nasal lavages were performedwith 10 mL of lactated Ringer’s solution.

Sample processing

All recovered nasal lavages were kept on ice until centri-fuged at 2500× g for 15 min at 48C. Aliquots of supernatantswere treated as appropriate for each mediator or protein tobe measured and were stored at¹808C until assayed. Whentwo lavages were performed, supernatants from the firstlavage were used for mediator and protein assays, while cellpellets from both lavages were combined and resuspendedfor cell counts. Aliquots of supernatants to be assayed foralbumin, lysozyme or tryptase were frozen with no prereat-ment. For aliquots of supernatants to be assayed for hista-mine, perchloric acid was added to a final concentration of2%, samples were centrifuged and the supernatant retainedfor assay. Aliquots for assay of prostaglandin D2 (PGD2)were made 80% in ethanol and incubated on ice for 10 minbefore centrifugation. Supernatants were dried undervacuum using a Speed-Vac Concentrator (Savant Instru-ments, Inc., Hicksville, NY, USA) and reconstituted inbuffer for assay. The remainder of each supernatant wasconcentrated<10-fold by placing the fluid in dialysis bagswith a 2000 molecular weight cut-off (Spectrum, Houston,TX, USA) and immersing them in a slurry of polyethyleneglycol 8000 (Sigma, St Louis, MO, USA), as previouslydescribed [10]. The concentrated supernatant was thendivided into aliquots for the assessment of cytokine content.

Mediator and protein assays

Histamine, PGD2 and tryptase were measured in lavagesobtained during the acute response to allergen challenge asindices of mast cell activation. We have previously shownthat, while histamine is produced during both the early andlate responses to allergen challenge, tryptase and PGD2 arenot produced during the late response [11]. Accordingly,histamine was measured in all lavages, but tryptase andPGD2 were not measured in lavages obtained at 4, 8 and24 h after challenge.

Levels of histamine and PGD2 were measured, aspreviously described, using an automated fluorometrictechnique [12] and by radioimmunoassay [13], respectively.

Tryptase was measured using a specific ELISA method.A previously described, rabbit polyclonal antibody tohuman tryptase [14] was coated on 96-well microtiterplates at a dilution of 1/1500 and incubated at 48C over-night. Plates were then washed and 1% sheep serum wasadded to each well to block non-specific binding sites. Afterwashing, a 100mL volume of tryptase standard, or samplesat various dilutions, were added to the plate and incubated at378C for 90 min. The plates were washed again and a100mL aliquot of a 1/4000 dilution of a biotinylatedpreparation of the same rabbit antibody to human tryptasewas added to each well and incubated at 378C for 90 min.After washing, a 1/1000 dilution of a 1 mg/mL solution of

streptavidin conjugated to horseradish peroxidase (Sigma,St Louis, MO, USA) was added to each well and incubatedat room temperature for 30 min. The plates were washed andthen developed using 0.4 mg/mLo-phenylenediamine dihy-drochloride (Sigma, St Louis, MO, USA) as the substrate,and the absorbance was read at 490 nm. The standard curvefor the ELISA ranged from 0.15 to 40 ng/mL.

Human serum albumin was used as an index of vascularpermeability and was measured using a specific ELISA,sensitive to 1 ng/mL of albumin as previously described [15].

Lysozyme was measured, as an index of serous glandularsecretion, in lavages obtained during the acute response toallergen challenge by specific ELISA as previouslydescribed [16]. Lysozyme was not measured in lavagesobtained at 4, 8, and 24 h after challenge because infiltratingneutrophils can also release this protein, making datainterpretation impossible.

Measurements of IL-3 and IL-8 were performed byELISA. In each case, polyclonal antibodies were producedby immunization of rabbits with purified, recombinantcytokine (Upstate Biotechnology, Lake Placid, NY, USA).An IgG fraction of each antibody was produced and aportion of the antibody was biotinylated using biotinyl-«-amino-caproic acid N-hydroxysuccinimide ester (Calbio-chem, La Jolla, CA, USA). For each cytokine, the ELISAprotocol was identical to that described above for tryptaseexcept that the dilutions of antibodies used to coat plateswere 1/500 for IL-3 and 1/1000 for IL-8, and biotinylatedantibodies were used at dilutions of 1/6000 for IL-3 and1/1200 for IL-8. The standard curve for the ELISA for IL-3ranged from 31 to 4000 pg/mL, while the standard curverange for IL-8 was from 59 to 15 000 pg/mL. Both assaysshowed<1% cross reactivity against a panel of othercytokines.

Interleukin-5 was measured by ELISA using reagents thatwere generously provided by Dr Richard Cook (SmithKlineBeecham, King of Prussia, PA, USA). Wells of microtiterplates were coated with a 1/1500 dilution of a monoclonalantibody to human IL-5 by overnight incubation at 48C.Plates were aspirated and buffer containing 1% bovineserum albumin (Sigma, St Louis, MO, USA) was added toeach well to block non-specific binding sites. After washing,IL-5 standard, or samples at various dilutions, were added tothe plate and incubated at 378C for 2 h. The plates werewashed again and a dilution of 1/200 of biotinylated rabbitpolyclonal antibody to human IL-5 was added to each welland incubated at 378C for 2 h. After washing, a 1/20 000dilution of a 1 mg/mL solution of streptavidin conjugated tohorseradish peroxidase (Sigma, St Louis, MO, USA) wasadded to each well and incubated at 378C for 30 min. Theplates were washed and then developed using 0.1 mg/mL3,30,5,50-tetramethylbenzidine (Sigma, St Louis, MO, USA)as the substrate, and the absorbance was read at 450 nm.

870 D. Proudet al.


The standard curve for the ELISA ranged from 3.9 to1000 pg/mL.

Cell counts

Cell pellets from lavages were re-suspended in 1 mL lactatedRinger’s solution and an aliquot was used in a haemocyto-meter chamber (Fisher, Pittsburgh, PA, USA) to determinetotal white blood cell counts. Samples that contained> 50 000 total white blood cells were again centrifuged,pellets were resuspended at 500 000 cells/mL, and 200mLaliquots (50,000–100 000 cells) were used to prepare cytos-pin slides (Shandon, Pittsburgh, PA, USA) for subsequentdifferential counts. From differential counts and total whiteblood cell counts, numbers of total eosinophils were calcu-lated. Cytospin slides were not prepared from samples with< 50 000 cells since prior experience shows that slides madeunder these conditions contain insufficient cells to permitdifferential counts.

Statistics

Because data were not normally distributed, values arepresented as medians together with values for 25th and75th percentiles, and non-parametric analyses were used tocompare the effects of salmeterol and placebo on theresponse to intranasal allergen challenge. For each para-meter measured during the acute phase of the allergenchallenge protocol, the primary data comparisons wereperformed using the sum of net changes from baseline. Toaccomplish this, values for a parameter observed after thediluent (baseline) challenge were subtracted from valuesrecorded after each dose of allergen, and the net values forthe three allergen doses were then added. Values obtained inthis manner from the salmeterol and placebo treatment dayswere compared using the Wilcoxon matched-pairs, signed-rank test. For data obtained at each late time point (4, 8, and

24 h), values were also expressed as net increase abovebaseline, and effects of treatment were again analysed usingthe Wilcoxon matched-pairs, signed-rank test. Significancewas assumed for values ofP# 0.05.

Results

Administration of a single intranasal dose of salmeterol didnot significantly affect allergen-induced sneezing or totalsymptom scores at any time during the study protocol(Table 1). Salmeterol also had no effect on the subjectiveevaluation of nasal congestion, when this was considered asan individual symptom (not shown).

Analysis of mediators and proteins measured during theacute allergen challenge portion of the protocol showed thatsalmeterol treatment did not significantly inhibit productionof any of the mast cell mediators, histamine, tryptase orPGD2 when expressed as sum of net changes from baseline.Although median concentrations of all three mediators weresomewhat lower after the 100 PNU allergen dose forsubjects on salmeterol than on placebo (Fig. 2), separateanalysis of these data showed that these differences did notachieve significance for any of the three mediators, eitherwhen assessed as increase above diluent or as absolutevalues. Salmeterol treatment also had no significant effecton release of the serous glandular marker, lysozyme (Fig.3A). By contrast, however, salmeterol inhibited allergen-induced increases in vascular permeability, as assessed bymeasurement of albumin levels (Fig. 3B). Albumin concen-trations were significantly lower (P< 0.02) when comparedas sum of net changes from baseline. Subsequent analysis asincrease above diluent at each data point indicated thatsignificant (P< 0.02 in each case) inhibition was observed atboth the 100 PNU and 1000 PNU allergen doses.

When data obtained at the late time points after allergenchallenge were analysed, there was, again, no effect ofsalmeterol treatment on histamine levels. Concentrations



Table 1.Allergen-induced sneezes and total symptom scores in subjects on salmeterol vs placebo (median [25th,75th percentiles])

Sneezes Symptom scores

Protocol time Placebo Salmeterol Placebo Salmeterol

Post diluent 0 [0, 0.5] 0 [0,0] 1.0 [0, 2.0] 0.5 [0, 1.5]Post 10 AU 3.0 [0, 10.0] 1.0 [0, 7.0] 3.5 [2.5, 5.5] 2.5 [0.5, 5.5]Post 100 AU 8.0 [2.0, 15.5] 9.0 [3.0, 16.0] 7.5 [4.5, 11.0] 9.0 [4.5, 12.5]Post 1000 AU 11.5 [7.5, 15.5] 12.0 [7.0, 16.0] 12.0 [10.5, 16.0] 12.5 [7.0, 14.5]4 h post antigen 2.0 [1.0, 4.5] 2.0 [0.5, 3.5] 6.0 [3.5, 8.0] 4.5 [3.0, 8.0]8 h post antigen 1.5 [0, 4.0] 0.5 [0, 4.0] 4.0 [3.0, 6.5] 4.0 [1.5, 5.5]24 h post antigen 4.0 [2.0, 10.5] 2.5 [1.0, 9.5] 3.0 [2.0, 8.5] 3.0 [2.0, 4.0]

of albumin measured at 4 h after allergen challenge werestatistically significantly lower (P< 0.05) on salmeterol thanon placebo. It was not possible to evaluate the effects ofsalmeterol on allergen-induced increases in cytokine levelsbecause no consistent, significant increases in IL-3, IL-5 orIL-8 were observed at any time point during the study.Finally, although allergen induced significant increases innumbers of eosinophils, and total white blood cells, in nasallavages, salmeterol treatment had no inhibitory effect onallergen-induced recruitment of inflammatory cells: no sig-nificant differences between salmeterol and placebo treat-ment in the increases in either eosinophil or total whiteblood cell counts above diluent levels were seen at any timepoint during the protocol (Table 2).

Discussion

The nasal challenge model used in this study was selected

because it permits the easy and objective evaluation ofseveral components of airway responses to allergen provo-cation, rather than as a method of evaluating the therapeuticefficacy of salmeterol. Indeed, given that there is no pre-cedent for any utility of b2-adrenergic agonists as aneffective therapy for allergic rhinitis, the lack of effect ofsalmeterol on nasal symptoms in the current study was notsurprising. It has been reported that high dose administra-tion of terbutaline, repeated before each of several challengedoses of allergen, led to a significant but extremely modestinhibition of symptoms induced only by the highest allergendose. This was due to a slight reduction in nasal blockage[17]. Given that, while the capacitance vessels that regulatenasal patency are under adrenergic control, they are sen-sitive primarily to a-adrenergic receptor stimulation [18],any major effect ofb2-adrenergic agonists on congestionwould be unlikely. The inability of salmeterol to modifynasal congestion in the present study therefore was notinexpected, particularly given the fact that subjects hadalready been premedicated with thea2-adrenergic agonist,oxymetazoline, as part of our protocol.

872 D. Proudet al.


Fig. 3.Effects of salmeterol treatment on serous glandular secretionand vascular permeability during the acute response to allergenchallenge. (a) lysozyme; (b) albumin. Heavy horizontal linesrepresent medians, while lower and upper limits of boxes represent25th and 75th percentile, respectively. *Values on salmeteroltreatment are significantly lower (P<0.02) than values afterplacebo treatment.

Fig. 2. Effects of salmeterol treatment on mast cell mediatorrelease during the acute response to allergen challenge. (a) hist-amine; (b) tryptase; (c) PGD2. Heavy horizontal lines representmedians, while lower and upper limits of boxes represent 25thand 75th percentile, respectively.

It has been well documented thatb-adrenergic agonistscan inhibit the IgE-mediated release of mediators fromhuman mast cells [19,20] and basophils [21]in vitro, andsalmeterol itself has been shown to inhibit the IgE-mediatedrelease of bronchospastic mediators from human lung tissue

[3]. Moreover, an inhaledb-adrenergic agonist has beenshown to inhibit plasma histamine release in response toantigen bronchoprovocationin vivo [22]. It was particularlysurprising therefore that salmeterol did not inhibit allergen-induced histamine release into nasal secretions at any timepoint during the current study. The failure of salmeterol toinhibit mast cell activation during the immediate response toallergen challenge was further supported, however, by theobservation that concentrations of the mast cell protease,tryptase, and of PGD2, the major cyclo-oxygenase productof human mast cells [23], were also unaffected by salme-terol pretreatment. The lack of any observed effect in thecurrent study is unlikely to be due to a failure of the drug toreach tissue mast cells, since salmeterol clearly penetratedtissues sufficiently to have effects on the nasal vasculature.One potential explanation would be if there was a hetero-geneity of nasal and pulmonary mast cells in terms of thedegree of their responsiveness tob-adrenergic agonists, butthis seems unlikely given the previous observation that highdose terbutaline inhibited allergen-induced tryptase releaseinto nasal secretions [17]. The difference between theobserved effects of terbutaline and our current data couldbe due to differences in dosage regimens, or to methodo-logical differences in collection of secretions (lavage vs a‘nasal pool’ technique). Regardless of the explanation, wehave been unable to confirm that a single dose of salmeterolmodulates allergen-induced mast cell activation in thehuman airways.

In addition to being innervated by cholinergic parasym-pathetic nerves, airway submucosal glands are known toexpress botha-adrenergic andb-adrenergic receptors. Inter-estingly, the serous cells of these glands contain morea-adrenergic thanb-adrenergic receptors, while the reversedistribution is seen on mucous cells [24]. Consistent withthis distribution, it has been shown that a robust secretion oflysozyme from serous glands can be induced by cholinergicor a-adrenergic stimulation bothin vitro andin vivo [25,26].Our observation that salmeterol did not modify serous



Table 2. Allergen-induced increases in eosinophil and total cell counts in subjects on salmeterol vs placebo(median [25th, 75th percentiles]).

Eosinophils (×10¹3) Total cells (×10¹5)

Protocol time Placebo Salmeterol Placebo Salmeterol

Post diluent 1.1 [0, 7.6] 0.4 [0, 3.5] 2.3 [0.5, 3.6] 0.6 [0.4, 1.6]Post 1000 AU 2.5 [0, 8.6] 1.5 [0, 17.5] 1.8 [0.5, 3.5] 1.0 [0.4, 2.7]4 h post antigen 17.7 [1.9, 81.7] 7.7 [4.1, 39.1] 4.0 [2.6, 5.9] 2.1 [1.3, 3.5]8 h post antigen 23.8 [9.2, 88.2] 36.4 [7.9, 124] 5.6 [2.5, 15.1] 7.7 [2.0, 14.5]24 h post antigen 23.2 [3.7, 153] 39.1 [8.6, 264] 4.6 [1.2, 12.6] 8.7 [1.8, 16.5]

Fig. 4. Effects of salmeterol treatment on histamine release andvascular permeability at the late time points after allergen challenge.(a) histamine; (b) albumin. Heavy horizontal lines representmedians, while lower and upper limits of boxes represent 25thand 75th percentile, respectively. *Values on salmeterol treat-ment are significantly lower (P< 0.05) than values after placebotreatment.

glandular secretion in the current study confirms earlierinvitro and in vivo demonstration of a lack of effect of otherb-adrenergic agonists [25,26]. We did not address theeffects of salmeterol on mucous cell secretion because wedid not have access to a reliable technique to measuremucous glycoproteins.

We were unable to evaluate the effects of salmeteroltreatment on production of IL-3, IL-5, members of the TH2

cytokine cluster, and IL-8, a chemokine produced byepithelial cells and other cells in the nasal mucosa, becauseallergen provocation did not induce significant increases ofany of these cytokines in our study population. Although ithas been reported that IL-5 increases after allergen chal-lenge of allergic subjects, modest increases were observedin only a subset of subjects [27]. These data are notinconsistent with our findings, in that two of our subjectshad modest increases in concentrations of IL-5 on placebo;obviously this did not permit any statistical evaluation ofdrug treatment effects. To our knowledge, this is the firstexamination of the effects of nasal challenge with allergenon IL-3 production. By contrast, there are conflicting priordata on the presence of IL-8 in allergic rhinitis. Althoughincreases were noted in allergic subjects after allergenchallenge [28], IL-8 concentrations have been reported todecrease during the allergy season [29]. Although we didnot detect increases in IL-8 in the present study, we do notbelieve that this is due to technical limitations, since thesame assay has detected increases in IL-8 after respiratorysyncytial virus infection in humans (D. Proud, unpublisheddata).

The most striking effect of salmeterol that was observedin our current study was the reduction of allergen inducedplasma efflux into the nasal airway, assessed by measure-ment of albumin in lavages. Ample precedent exists forb2-adrenergic agonists in general [30,31], and for salmeterolin particular [5], to alter permeability in the airways inanimal models. Indeed, our data are in good agreement withthe report that salmeterol can inhibit allergen-inducedplasma leakage in rat airways [32]. Moreover,b2-adrenergicagonists have been shown to inhibit both IgE-mediated [33]and histamine-induced [34] increases in vascular per-meability in human skin. Our data with salmeterol arealso in excellent agreement with those observed usinghigh dose terbutaline and nasal challenge with allergen,in that both drugs had no effect on baseline albumin levelsbut showed the most pronounced inhibition at the higherallergen doses.

It has been suggested that the ability ofb2-adrenergicagonists to inhibit plasma exudation in the airways may bean important mode of action of these drugs in asthmaticpatients [35,36]. Increased vascular permeability may leadto the generation of plasma-derived mediators in the airways[35] and has been proposed to contribute significantly to

the pathogenesis of asthma by causing luminal narrowing[37]. Moreover, luminal narrowing due to plasma exudationin asthmatic subjects could exaggerate increases in airwayconstriction induced by bronchoconstrictive mediators [38].

Finally, although it has been reported that inhaled salme-terol can inhibit allergen-induced leucocyte adhesion in rats[32], as well as lipopolysaccharide induced accumulation ofneutrophils, and infiltration of eosinophils in response toplatelet activating factor, in guinea-pig airways [5], we wereunable to demonstrate any inhibitory effects of salmeterolon allergen-induced infiltration of eosinophils or total whiteblood cells into the human nasal mucosa. Although wecannot rule out that an inhibition of cellular influx couldconceivably have occurred if higher doses of salmeterolwere used, it is, perhaps, more likely that our data can bebest explained on the basis of species differences. Certainly,there is no indication to date that salmeterol reducesinflammatory cell populations in the airways of asthmapatients.

In summary, we have shown that salmeterol inhibitsantigen-induced increases in vascular permeability butdoes not alter mast cell activation, serous glandular secre-tion or cellular infiltration. The ability of salmeterol toinhibit antigen-induced vascular permeability may contri-bute to its therapeutic efficacy in asthma but directin vivostudies on the effects ofb2-adrenergic agonists on per-meability in the human lower airways are necessary todefinitively establish this.

Acknowledgements

This work was supported by a grant-in-aid from Glaxo-Wellcome.

References

1 Castle W, Fuller R, Hall J, Palmer J. Serevent nationwidesurveillance study:comparison of salmeterol with salbutamolin asthmatic patients who require regular bronchodilator treat-ment. BMJ 1993; 306:1034–7.

2 Ullman A, Svedmyr N. Salmeterol, a new long acting inhaled2 adrenoreceptor agonist: comparison with salbutamol in adultasthmatic patients. Thorax 1988; 43:674–8.

3 Butchers PR, Vardey CJ, Johnson M. Salmeterol: a potent andlong-acting inhibitor of inflammatory mediator release fromhuman lung. Br J Pharmacol 1991; 104:672–6.

4 Twentyman OP, Finnerty JP, Harris A, Palmer J, Holgate ST.Protection against allergen-induced asthma by salmeterol.Lancet 1990; 336:1338–42.

5 Whelan CJ, Johnson M. Inhibition by salmeterol of increasedvascular permeability and granulocyte accumulation in guinea-pig lung and skin. Br J Pharmacol 1992; 105:831–8.

6 Naclerio RM, Meier HC, Kagey-Sobotka A et al. Mediatorrelease after nasal airway challenge with allergen. Am RevRespir Dis 1983; 128:597–602.

874 D. Proudet al.


7 Majchel AM, Baroody F, Kagey-Sobotka A, Lichtenstein LM,Naclerio RM. Effect of oxymetazoline on the early response tonasal challenge with antigen. J Allergy Clin Immunol 1993;92:767–70.

8 Churchill L, Pongracic JA, Reynolds CJ, Naclerio RM, ProudD. Pharmacology of nasal provocation with bradykinin: studiesof tachyphylaxis, cyclooxygenase inhibition, alpha-adrenergicstimulation, and receptor subtype. Int Arch Allergy ApplImmunol 1991; 95:322–31.

9 Svensson C, Pipkorn U, Alkner U, Baumgarten CR, PerssonCGA. Topical vasoconstrictor (oxymetazoline) does not affecthistamine-induced mucosal exudation of plasma in humannasal airways. Clin Exp Allergy 1992; 22:411–6.

10 Proud D, Gwaltney JM Jr, Hendley JO, Dinarello CA, Gillis S,Schleimer RP. Increased levels of interleukin-1 are detected innasal secretions of volunteers during experimental rhinoviruscolds. J Infect Dis 1994; 169:1007–13.

11 Naclerio RM, Hubbard W, Lichtenstein LM, Kagey-SobotkaA, Proud D. Origin of late phase histamine release. J AllergyClin Immunol 1996; 98:721–3.

12 Siraganian RP. Refinements in the automated fluorometrichistamine analysis system. J Immunol Meth 1975; 7:283–90.

13 Schulman ES, Newball HH, Demers LM, Fitzpatrick FA,Adkinson NF Jr. Anaphylactic release of thromboxane A2,prostaglandin D2, and prostacyclin from human lung paren-chyma. Am Rev Respir Dis 1981; 124:402–6.

14 Proud D, Bailey GS, Naclerio RM et al. Tryptase and histamineas markers to evaluate mast cell activation during the responsesto nasal challenge with allergen, cold, dry air, and hyperosmo-lar solutions. J Allergy Clin Immunol 1992; 89:1098–110.

15 Riccio MM, Proud D. Evidence that enhanced nasal reactivityto bradykinin in patients with symptomatic allergy is mediatedby neural reflexes. J Allergy Clin Immunol 1996; 97:1252–63.

16 Riccio MM, Reynolds CJ, Hay DW, Proud D. Effects ofintranasal administration of endothelin-1 to allergic and non-allergic individuals. Am J Respir Crit Care Med 1995;152:1757–64.

17 Svensson C, Greiff L, Andersson M, Alkner U, Gro¨nnenbergR, Persson CGA. Antiallergic actions of high topical doses ofterbutaline in human nasal airways. Allergy 1995; 50:884–90.

18 Bende M, Loth S. Vascular effects of topical oxymetazoline onhuman nasal mucosa. J Laryngol Otol 1986; 100:285–8.

19 Peters SP, Schulman ES, Schleimer RP et al. Dispersed humanlung mast cells. Pharmacologic aspects and comparison withhuman lung tissue fragments. Am Rev Respir Dis 1982;126:1034–9.

20 Church MK, Hiroi J. Inhibition of IgE-dependent histaminerelease from human dispersed lung mast cells by anti-allergicdrugs and salbutamol. Br J Pharmacol 1987; 90:421–9.

21 Lichtenstein LM, Margolis S. Histamine release in vitro:inhibition by catecholamines and methylxanthines. Science1968; 161:902–3.

22 Howarth PH, Durham SR, Lee TH et al. Influence of albuterol,cromolyn sodium and ipratropium bromide on the airway andcirculating mediator response to allergen bronchial provoca-tion in asthma. Am Rev Respir Dis 1985; 132:986–92.

23 Peters SP, MacGlashan DW Jr, Schulman ES et al. Arachidonicacid metabolism in purified human lung mast cells. J Immunol1984; 132:1972–9.

24 Barnes PJ, Basbaum CB. Mapping of adrenergic receptors inthe trachea by autoradiography. Exp Lung Res 1983; 5:183–92.

25 Tom-Moy M, Basbaum CB, Nadel JA. Localization andrelease of lysozyme from ferret trachea: effects of adrenergicand cholinergic drugs. Cell Tissue Res 1983; 228:549–62.

26 Mullol J, Raphael GD, Lundgren JD et al. Comparison ofhuman nasal mucosal secretion in vivo and in vitro. J AllergyClin Immunol 1992; 89:584–92.

27 Sim TC, Grant JA, Hilsmeier KA, Fukuda Y, Alam R. Pro-inflammatory cytokines in nasal secretions of allergic subjectsafter challenge. Am J Respir Crit Care Med 1994; 149:339–44.

28 Sim TC, Reece LM, Hilsmeier KA, Grant JA, Alam R.Secretion of chemokines and other cytokines in allergen-induced nasal responses: inhibition by topical steroid treat-ment. Am J Respir Crit Care Med 1995; 152:927–33.

29 Kuna P, Lazarovich M, Kaplan AP. Chemokines in seasonalallergic rhinitis. J Allergy Clin Immunol 1996; 97:104–12.

30 Persson CGA, Ekman M, Erjefa¨lt I. Vascular anti-permeabilityeffects of b-receptor agonists and theophylline in the lung.Acta Pharmacol Toxicol 1979; 44:216–20.

31 Erjefalt I, Persson CGA. Long duration and high potency ofantiexudative effects of formoterol in guinea-pig tracheobron-chial airways. Am Rev Respir Dis 1991; 144:788–91.

32 Bolton PB, LeFevre P, McDonald DM. Salmeterol reducesearly- and late-phase plasma leakage and leukocyte adhesionin rat airways. Am J Respir Crit Care Med 1997; 155:1428–35.

33 Gronneberg R, Zetterstro¨m O. Inhibition of anti-IgE inducedskin response in normals by formoterol, a newb2-adreno-receptor agonist and terbutaline. Allergy 1990; 45:340–6.

34 Basran GS, Paul W, Morley J, Turner-Warwick M. Adreno-receptor agonist inhibition of the histamine induced cutaneousresponse in man. Br J Dermatol 1982; 107 (Suppl. 23):140–2.

35 Persson CGA. Role of plasma exudation in asthmatic airways.Lancet 1986; ii:1126–9.

36 Erjefalt I, Persson CGA. Pharmacologic control of plasmaexudation into tracheobronchial airways. Am Rev Respir Dis1991; 143:1008–14.

37 Pare PD, Godden DJ. Airway edema. In: Crystal RG, West JB,Barnes PJ, Chernial NS, Weibel ER, eds The Lung: ScientificFoundations. New York: Raven Press, 1991:987–94.

38 Yager D, Butler JP, Bastacky J, Israel G, Smith G, Drazen JM.Amplification of airway constriction due to liquid filling orairway interstices. J Appl Physiol 1989; 66:2873–84.



intranasal salmeterol inhibits allergen-induced vascular permeability but not mast cell activation...

Documents