496

Mechanism

s of allergy

Eosinophil cationic protein alterspulmonary surfactant structure andfunction in asthma

Jens M. Hohlfeld, MD,a,c Andreas Schmiedl, MD,b Veit J. Erpenbeck, MD, PhD,a,c

Per Venge, MD,d and Norbert Krug, MDc Hannover, Germany, and Uppsala, Sweden

Background: Impaired surfactant function has been demon-strated in patients with asthma. Inhibitory proteins originatingfrom plasma or inflammatory mediators are good candidatesto contribute to this dysfunction. Eosinophils are potent effec-tor cells in asthma, which, on activation, release inflammatorymediators, especially reactive granula proteins such aseosinophil cationic protein (ECP).Objective: Because the potential role of ECP in the inhibition ofsurfactant function is not known, we tested the hypothesis ofwhether ECP levels in bronchoalveolar lavage fluid (BALF) ofpatients with asthma after segmental allergen provocation corre-late to surfactant dysfunction. Furthermore, we tested the effectof purified ECP on surfactant function and structure in vitro.Methods: Surfactant isolated from BALF of asthmatic patientswas assessed for biophysical function with the Pulsating Bub-ble Surfactometer and the Capillary Surfactometer and corre-lated to ECP levels. Purified ECP and plasma proteins at vari-ous concentrations were incubated with natural surfactant.Surfactant function was studied with the Capillary Surfac-tometer, and surfactant structure was determined by electronmicroscopy.Results: ECP is elevated in BALF from patients with asthmaafter allergen challenge compared with baseline. ECP levelsafter allergen challenge correlate well to surfactant dysfunc-tion. In vitro, ECP induces a concentration-dependent inhibi-tion of surfactant function that can be inhibited by antibodiesagainst ECP. ECP is more potent compared with albumin orfibrinogen. Finally, ECP induces severe ultrastructuralchanges to surfactant vesicles that are more pronounced thanchanges induced by either fibrinogen or albumin.Conclusions: ECP contributes to surfactant dysfunction inasthma, which in turn could lead to airway obstruction. (JAllergy Clin Immunol 2004;113:496-502.)

Key words: Asthma, airway obstruction, eosinophils, surfactantfunction, surfactant ultrastructure

Pulmonary surfactant covers the surface of alveoli andairways by forming a surface-active layer at the air-liquidinterface. Reduction of surface tension in the alveoli pre-vents end-expiratory alveolar collapse and thus allows fora normal ventilation of the entire lung. In the airways, sur-factant has also been suggested to play a role because itguarantees airway patency as the result of its high surfacepressure.1,2 Impaired surfactant function might thus con-tribute to airway narrowing and closure. Recently, we andothers have shown that surfactant function is disturbed inpatients with asthma during exacerbation and after seg-mental allergen provocation.3-5 In asthma, several mecha-nisms might be involved in the inhibition of surfactantfunction. It is well known that various proteins, such asalbumin, fibrinogen, and hemoglobin, are potent surfac-tant inhibitors.6 We have suggested that in the acute asth-ma attack, plasma proteins entering the airways inhibitproper surfactant function. This hypothesis is supportedby the finding of increased phosphatidylcholine speciescontaining linoleic acid in bronchoalveolar lavage fluid(BALF) of asthmatic patients after allergen challenge.7

These molecular species were characteristic of plasmaphosphatidylcholine, suggesting infiltration of plasmalipoproteins. Further evidence comes from the fact thatdisturbed surfactant function could be restored by wash-ing the BALF surfactant pellet, a procedure that removeswater-soluble inhibitors such as plasma proteins.3

Non–plasma-derived proteins might also play a role inthe inhibition of surfactant function. In this respect,eosinophil granule proteins are major candidates becauseeosinophils are a predominant cell type found in the asth-matic inflammatory process. Activated eosinophilsrelease various cytotoxic proteins such as eosinophilcationic protein (ECP), arachidonic acid metabolites, andoxygen-derived radicals.8 This causes damage to the air-way epithelium associated with bronchial hyperrespon-siveness and episodic bronchial obstruction, the clinicalcharacteristics of asthma. Elevated levels of ECP can be

From the Departments of aRespiratory Medicine and bAnatomy, HannoverMedical School, Hannover, Germany; the cDepartment of Immunology,Allergology, and Clinical Inhalation, Fraunhofer Institute of Toxicologyand Experimental Medicine, Hannover, Germany; and the dDepartment ofClinical Chemistry, University Hospital Uppsala, Sweden.

Received for publication June 26, 2003; revised November 18, 2003; accept-ed for publication December 5, 2003.

Reprint requests: Jens M. Hohlfeld, MD, Department of Respiratory Medi-cine, Hannover Medical School, Carl-Neuberg-Str 1, D-30625 Hannover,Germany.

Supported by a grant of the Deutsche Forschungsgemeinschaft (SFB587/B8).

0091-6749/$30.00© 2004 American Academy of Allergy, Asthma and Immunologydoi:10.1016/j.jaci.2003.12.008

Abbreviations usedBALF: Bronchoalveolar lavage fluid

CS: Capillary SurfactometerECP: Eosinophil cationic protein

LA: Large surfactant aggregatesPC20: Concentration of histamine that causes a 20% fall in

FEV1ULV: Unilamellar vesiclesγmin: Surface tension at minimal bubble size

J ALLERGY CLIN IMMUNOL

VOLUME 113, NUMBER 3

Hohlfeld et al 497

Mec

hani

sms

of a

llerg

y

found both in sputum and in BALF from patients withasthma.9,10 In the airways, ECP gets into close contactwith pulmonary surfactant present in the hypophase of theairway lining layer. Therefore, we hypothesized thateosinophil cationic protein released into the airways alterssurfactant function. To investigate this hypothesis, weasked whether surfactant dysfunction induced by segmen-tal allergen challenge correlates to ECP levels in BALFfrom patients with asthma. In addition, we conducted anin vitro study to investigate the effect of ECP, albumin,and fibrinogen on biophysical surfactant functionassessed by a Capillary Surfactometer (CS), which simu-lates the architecture of small airways. Furthermore, weassessed the effect of these proteins on the ultrastructuralmorphology of natural surfactant by electron microscopy.

METHODS

Segmental allergen challenge

Levels of ECP, albumin, and fibrinogen of BALF from patientswith asthma after local allergen challenge were correlated to the bio-physical properties of surfactant isolated from the same sample. Dataon surfactant activity have recently been published by our groupelsewhere.3 What follows is a brief summary of the study design.

Fourteen patients with mild intermittent allergic asthma accord-ing to the Heart, Lung, and Blood Institute of the National Institutesof Health11 and 9 healthy volunteers were enrolled in the bron-choscopy study. Baseline characteristics of all study subjects (TableI) were determined by skin prick testing, spirometry, serum IgE lev-els, and bronchial hyperresponsiveness, which was determined by amodified bronchoprovocation test with histamine, as described else-where.12 Briefly, aerosols of histamine were inhaled at 5-minuteintervals in doubling concentrations (0.03 mg/mL up to 16 mg/mL)until a fall in FEV1 of 20% was reached or when the highest doseof histamine had been given. Normal subjects were required to havea PC20 of >8 mg/mL. In contrast, a PC20 of <8 mg/mL was not aprerequisite in asthmatic subjects because some of them were test-ed out of their season while asymptomatic.12 None of the patientshad used glucocorticosteroids, sodium cromoglycate, leukotrienemodifiers, or theophylline for at least 6 weeks. All study subjectswere nonsmokers, and none had acute bronchitis during the 4-weekperiod preceding the challenge. The study was approved by theEthics Committee of Hannover Medical School, and informed con-sent was obtained from each person in the study.

Segmental allergen challenge was performed as previouslydescribed.3,13 The allergen extract used for segmental allergen chal-lenge (mixed grass pollen or D pteronyssinus; Abelló) was thatwhich produced the largest wheal response on skin-prick testing,and the chosen concentration for endobronchial challenge was one-tenth the dilution in saline that elicited a 3-mm–diameter skin-wheal response. The instilled amount of antigen was 0.01 to 0.1 µgfor mixed grass pollen (a mixture of the major allergens Dac g5, Fesp5, Lol p5, Phl p5, and Poa p5) and 0.6 µg for D pteronyssinus (amixture of the major allergens Der p1 and Der p2). Healthy controlsubjects received an allergen challenge with a fixed dose of eithergrass pollen allergen (0.1 µg) or house dust mite allergen (0.6 µg)in randomized fashion. After performing a baseline BAL in the infe-rior lingular bronchus, 10 mL of saline solution was instilled intothe superior lingular bronchus as a control. Allergen solution (10mL) was then instilled into the medial segment of the middle lobe.After 24 hours, subjects (all patients with asthma and 5 control sub-jects) were given rebronchoscopy, and the superior lingularbronchus and the medial middle lobe were lavaged.

Bronchoalveolar lavage

The fiberoptic bronchoscope was wedged into the appropriatebronchus, and lavage was done with warm saline solution in 6aliquots of 20 mL. After instillation, each aliquot was aspirated withgentle suction. The first 20-mL aliquot was sampled separately. Theremaining BALF was pooled, and the recovered volume was record-ed (“recovery”). Specimens of 5 mL were sent for total count ofcells and for differential cell counts. The remaining fluid was fil-tered through sterile gauze and centrifuged at 250g for 10 minutesto obtain a cell-free supernatant, which was stored at –28°C untilfurther analysis. All further measurements were performed blindly.

Determination of eosinophil granule protein

and plasma proteins

Aliquots of the cell-free supernatant were used for determinationof protein concentrations according to Lowry et al14 and ofeosinophil cationic protein by commercially available Uni-CAP-ECP system (Pharmacia Diagnostics AB, Uppsala, Sweden). Fi-brinogen was measured by ELISA, with the use of a matched-pairantibody set (Affinity Biologicals, Ancaster, Canada). A fibrinogenstandard was obtained from Enzyme Research Laboratories (SouthBend, Ind). For the measurement, BALF samples were diluted 1:5or 1:10. Albumin was determined by nephelometry. Antiserum andstandards for albumin were obtained from Dade Behring, and sam-ples were analyzed on a Dade Behring nephelometer according tothe manufacturer’s instructions (Dade Behring, Marburg, Ger-

TABLE I. Clinical characteristics of the study subjects

FEV1 IgE

Age Sex (% predicted) PC20 (mg/mL)

Control subjects 33 F 103 10.2 6232 M 95 11.9 523 M 103 10.5 5227 M 101 20.0 723 M 104 25.4 725 M 108 11.9 324 M 103 11.3 928 M 96 16.1 833 F 104 20.1 8

Mean 27.6 ± 1.4 2 F/7 M 102 ± 1 14.5 10

Asthmatic 24 F 106 11.2 50subjects 22 M 91 12.4 310

29 M 101 5.2 39022 F 100 5.6 41527 M 109 2.9 32330 F 70 0.3 33924 F 90 2.6 72828 M 78 1.5 6428 M 99 8.4 11026 M 96 3.6 7324 M 90 8.5 16826 M 119 4.1 19423 F 79 0.5 9622 M 115 13.0 4028 F 109 1.0 78

Mean 25.5 ± 0.7 6 F/9 M 97 ± 4 3.4* 159*

FEV1 was measured before the first bronchoscopy. For age and FEV1, mean± SEM are given. Bronchial hyperresponsiveness (PC20) was defined as thehistamine concentration leading to a 20% drop of baseline FEV1. For PC20and IgE, geometric means are shown.*P < .01 compared with control subjects.

498 Hohlfeld et al J ALLERGY CLIN IMMUNOL

MARCH 2004

Mechanism

s of allergy

many). All assays were performed in duplicate, and the mean valuewas reported.

Experiments with the Capillary

Surfactometer

The basic principle of the Capillary Surfactometer has beenextensively described.1,15,16 Briefly, this instrument simulates themorphology and function of a terminal conducting airway with aglass capillary that in a short section is particularly narrow, with anID of 0.25 mm. A small volume (0.5 µL) of the liquid to be evaluat-ed is deposited into this section. Pressure is raised at one end of thecapillary, extruding the liquid from the narrow section. Pressure iszero if the capillary is open for free airflow. Pressure is recorded for120 seconds, and a computer calculates the percentage of that timethat the pressure is zero—that is, the capillary is open (Open in %).

To study the function of surfactant isolated from BALF of asth-matic subjects, a volume of 500 µL of the cell-free BALF was cen-trifuged at 4°C and 40,000g for 1 hour. After removal of 400 µLsupernatant, the remaining 100 µL, containing the large surfactantaggregates (LA), was exposed to vacuum in a small test tube.Through a very fine glass capillary, dry nitrogen was sucked into thetest tube to completely dehydrate the sample. The sample was thenresuspended in 25 µL of the supernatant and carefully stirred. Fivealiquots of 0.5 µL each were studied with the CS, and the meanvalue was reported.

To study the effect of ECP, albumin, and fibrinogen on surfactantfunction in vitro, we used a commercially available natural bovinesurfactant (Alveofact). Alveofact was diluted with saline and usedat a final phopholipid concentration of 0.1 mg/mL. Albumin (50 to10,000 µg/mL), fibrinogen (10 to 5000 µg/mL), and ECP (10 to 500µg/mL) were diluted in saline and admixed the concentrations indi-cated. In an additional set of experiments, the specificity of ECP onsurfactant function was investigated by adding polyclonal anti-ECP(1:1000) to ECP (100 µg/mL) 5 minutes before admixture withAlveofact. Three aliquots of 0.5 µL each were studied with the CS,and the mean value was reported. The experiments were repeatedfor each concentration on 4 to 6 different occasions (n = 4 to 6).

Experiments with the Pulsating Bubble Sur-

factometer

Surface activity of BALF surfactant was additionally measuredwith a Pulsating Bubble Surfactometer (Electronetics, Buffalo,NY). Surfactant was isolated from cell-free BALF by centrifugationat 48,000g for 60 minutes at 4°C. The pellet contained the LA. Thesupernatant was removed, and the LA pellet was resuspended inRinger solution. Forty microliters of the LA suspension, which hadbeen given a phospholipid concentration of 1 mg/mL, was used forfilling the sample chamber of the Pulsating Bubble Surfactometerwith a micropipet. The surface tension used for statistical analysisof this study was the value at minimal bubble size, γmin, registeredafter 5 minutes of pulsation at a rate of 20 cycles per minute and ata temperature of 37°C. All analog data were digitalized and record-ed by computer.

Electron microscopy

Alveofact at a concentration of 0.2 mg/mL was mixed 1:1 witheither ECP, fibrinogen, or albumin, respectively, at final concentra-tions of 100 µg/mL for 2 hours. Thereafter, fixation solution (3%glutaraldehyde and 3% paraformaldehyde in 0.2 mol/L HEPESbuffer) was added to the mixtures (1:1). After fixation for 2 hours,the suspensions were centrifuged at 40,000g for 4.5 hours. Theresulting pellets were rinsed with 0.1 mol/L cacodylate buffer (3 ×10 minutes). Stabilization of predominantly unsaturated lipids wasachieved by postfixation with osmium tetroxid (1% OsO4 in 0.1

mol/L cacodylate buffer). After rinsing in cacodylate buffer (3 × 10minutes) and distilled water (2 × 10 minutes), the pellets werestained en bloc overnight with a mixture of equal portions of uranylacetate and water (half-saturated aqueous uranyl acetate solution[1:1]). After dehydration through an ascending series of acetone, thepellets were embedded in epon. Ultrathin sections (70 nm) were cutand stained with lead citrate and uranyl acetate. Qualitative analysisof surfactant morphology was carried out on stained ultrathin sec-tion with the use of an electron microscope (EM 10, Zeiss,Oberkochen, Germany).

Quantitative determination of the volume density of unilamellarvesicles was performed by point and intersection, counting accord-ing to Weibel,17 by using a coherent multipurpose test system with13 test lines and 26 test points. This stereologic method was carriedout on-line (SIT 66-X camera, DAGE-MIT Inc, Michigan City,Mich; connected to a computer), using an image software package(AnalySIS 2.0, Soft Imaging System, Münster, Germany) with anEM 902 (Zeiss, Oberkochen, Germany).

Materials

Alveofact, a lipid extract surfactant preparation from bovinelung lavage lacking SP-A and SP-D, was kindly donated byBoehringer Ingelheim, Biberach, Germany. For the in vitro assays,human serum albumin (approximately 99%) was purchased fromSigma (Deisenhofen, Germany), and human fibrinogen wasobtained from Enzyme Research Laboratories Inc (South Bend,Ind). ECP and polyclonal anti-ECP were generously provided byPer Venge, Uppsala, Sweden.

Statistics

Results are expressed as mean ± SEM. For statistical analysis ofECP data from the clinical study, diversity in means of the treatmentgroups (baseline, saline, and allergen) within the major studygroups (asthmatic and control subjects) was tested by nonparamet-ric Friedman test for paired comparison followed by Wilcoxon ranktest for individual comparison between treatment groups. Differ-ences between the major study groups were tested by Mann-Whit-ney U test.18 Correlation of ECP data with surfactant function wasdone by Spearman rank test. Statistical differences between meanvalues of the in vitro dose-response curves of the various proteinswere determined by Kruskal-Wallis test followed by Mann-WhitneyU test for individual comparison between treatment groups. P val-ues of <.05 were considered significant.

RESULTS

Effect of ECP on surfactant function

Segmental allergen provocation induced a markedeosinophilic inflammation. Eosinophil numbers in BALFwere 0.2 ± 0.1 × 106 at baseline; allergen challengeinduced a tremendous increase in eosinophil numbers(55.0 ± 34.3 × 106), but challenge with saline did not (2.0± 0.7 × 106). In healthy control subjects, eosinophilswere nominally absent at baseline and after saline andallergen challenge.

After allergen challenge, there was a marked increasein ECP levels from BALF of asthmatic subjects com-pared with baseline (Fig 1), whereas saline challenge didnot change levels of ECP. In contrast, saline challengecompared with baseline BALF from asthmatic subjectsinduced a significant increase of both fibrinogen (base-line: 429.0 ± 110.8 µg/L, saline: 877.3 ± 233.5 µg/L) andalbumin (baseline: 57.8 ± 8.4 mg/L, saline: 91.2 ± 14.4

J ALLERGY CLIN IMMUNOL

VOLUME 113, NUMBER 3

Hohlfeld et al 499

Mec

hani

sms

of a

llerg

y

mg/L). As for ECP, allergen challenge induced a markedand significant increase of both fibrinogen (3648 ± 1053µg/L) and albumin (980.7 ± 504.9 mg/L). Neither salinenor allergen induced significant changes of ECP, fibrino-gen, and albumin in healthy volunteers compared withbaseline.

ECP levels in BALF from asthmatic subjects afterallergen challenge correlated to surfactant dysfunctiondetermined with the Capillary Surfactometer (r = –0.63;P = .016; Fig 2, A). ECP levels also correlated to surfac-tant dysfunction measured with the Pulsating BubbleSurfactometer (r = 0.71; P = .004; Fig 2, B). There wasalso a very good correlation of surfactant dysfunction(γmin) with the BALF concentration of albumin (r = 0.91,P < .0001). However, BALF fibrinogen did not correlateto surfactant dysfunction (r = 0.17, P = .61).

Function of natural surfactant (Alveofact, 0.1 mg/mL)was assessed in the presence of various concentrations ofdifferent proteins by using the Capillary Surfactometer.The openness of the capillary was diminished at increas-ing protein concentrations for all proteins tested. Naturalsurfactant at 0.1 mg phospholipids/mL was functionallyactive with an openness of the test tube of 99.95% ±0.02%, whereas addition of the proteins decreased open-ness in a concentration-dependent manner (Fig 3). Half-maximal effect (potency) of the proteins to inhibit sur-factant function with a reduction of capillary openness to50% was achieved for ECP at a concentration of 110µg/mL, whereas fibrinogen (860 µg/mL) and albumin(2370 µg/mL) were less potent (Fig 3).

When ECP was preincubated with anti-ECP beforeadmixture with Alveofact, the inhibitory effect of ECP onsurfactant function was clearly inhibited. Openness ofthe capillary with ECP (100 µg/mL) alone was 32.8% ±5.9%, whereas preincubation of ECP with anti-ECPresulted in 86.2% ± 2.3% openness (P < .0001, Fig 4).Anti-ECP alone had no effect on openness of the capil-lary (96.4% ± 1.4%).

Effect of ECP on surfactant morphology

We used electron microscopy to visualize the effect ofthe different proteins on surfactant morphology. Naturalsurfactant (Alveofact) at a concentration of 0.1 mg phos-pholipids/mL mainly contained large multilamellar bod-ies with dense lamellae and very few unrolled lipid mem-branes (Fig 5, A). Albumin (100 µg/mL) had only littleeffect on surfactant ultrastructure. Multilamellar bodieswere slightly smaller, and the lipid layers of the multi-lamellar bodies showed a less dense order (Fig 5, B). Fi-brinogen (100 µg/mL) induced unrolling of multilamel-lar bodies, leading to deformation of the round-shapedvesicles and breaking up of lipid layers. Some smallermultilamellar vesicles and few unilamellar vesicles werealso present (Fig 5, C). In contrast, ECP (100 µg/mL)induced numerous unilamellar vesicles and small multil-amellar vesicles compared with albumin and fibrinogen.Interestingly, some of the unilamellar vesicles are presentwithin larger multilamellar structures. As albumin andfibrinogen, ECP also leads to unrolling of multilamellarlipid layers and deformation of round-shaped vesicles(Fig 5, D).

FIG 1. Levels of ECP in BALF from healthy control subjects andpatients with asthma at baseline (B) and after challenge witheither saline (S) or allergen (A). Results are given as mean ± SEM.**P < .001 compared with baseline.

FIG 2. Correlation of biophysical surfactant function with ECP lev-els in BALF from patients with asthma after segmental allergenprovocation. A, Functional surfactant data measured with theCapillary Surfactometer. Percentage of openness of the test capil-lary (Open in %) during 120 seconds is given. Surface tension atminimum bubble size (γmin) after 5 minutes of oscillation in thePulsating Bubble Surfactometer is shown in B.

500 Hohlfeld et al J ALLERGY CLIN IMMUNOL

MARCH 2004

Mechanism

s of allergy

In addition, we quantified the volume density of uni-lamellar vesicles (ULV) in the surfactant suspensionswithout and with the proteins. Albumin and fibrinogen ata protein concentration of 100 µg/mL induced only minor

changes, with volume densities of ULV of 3.2% ± 2.0%and 8.6% ± 3.3%, respectively, compared with Alveofactalone (1.2% ± 0.5%). In contrast, ECP resulted in amarked increase of ULV formation (31.2% ± 3.6%; Fig 6).

FIG 3. Effect of increasing concentrations of eosinophil cationicprotein (closed circles), fibrinogen (open squares), and albumin(open circles) on surfactant function measured with the CapillarySurfactometer as the percentage of openness of test capillariescontaining natural bovine surfactant (Alveofact) at a phospholipidconcentration of 0.1 mg/mL. Data are expressed as mean ± SEM.*P < .05 compared with albumin.

FIG 4. Effects of adding ECP alone (100 µg/mL), ECP (100 µg/mL)together with polyclonal antibodies against ECP (anti-ECP,1:1000), and anti-ECP alone to natural bovine surfactant (Alveo-fact) at a phospholipid concentration of 0.1 mg/mL on surfactantfunction measured with the Capillary Surfactometer as the per-centage of openness of test capillaries. Data are expressed asmean ± SEM. **P < .01 compared with control.

FIG 5. Ultrastructure of natural bovine surfactant at a phospholipid concentration of 0.1 mg/mL (A). Effectsof albumin (B), fibrinogen (C), and ECP (D) on surfactant structure is shown. All proteins were used at a pro-tein concentration of 100 µg/mL. Note numerous ULV induced by ECP (arrow, D).

J ALLERGY CLIN IMMUNOL

VOLUME 113, NUMBER 3

Hohlfeld et al 501

Mec

hani

sms

of a

llerg

y

DISCUSSION

It has been suggested that surfactant plays a role in theregulation of airway caliber and that dysfunction of sur-factant lining the airways contributes to airway obstruc-tion in asthma.19,20 In patients with asthma, allergenchallenge induces a dysfunction of surfactant isolatedfrom BALF.3,4 The exact mechanisms of surfactant dys-function in asthma have not been elucidated. Proteininhibition serves as one likely explanation. Besides plas-ma proteins such as albumin and fibrinogen that invadethe airways during the acute asthma attack, proteins syn-thesized or released locally during the allergiceosinophilic inflammation might contribute to surfactantdysfunction. Thus, we hypothesized that eosinophil gran-ule proteins induce alterations of pulmonary surfactant.

In this study, we present for the first time data demon-strating that ECP is a potent inhibitor of surfactant func-tion. In the clinical part of this study, we found elevatedlevels of ECP in BALF from asthmatic subjects after seg-mental allergen provocation (Fig 1). These results confirmrecent findings of increased ECP in BALF after allergenchallenge by either inhalation21 or segmental instilla-tion.22,23 Interestingly, ECP levels correlated with thedegree of surfactant dysfunction estimated either with theCapillary Surfactometer or with the Pulsating Bubble Sur-factometer (Fig 2). However, interpretation of this correla-tion as a causative relation is limited by the fact that thedegree of surfactant dysfunction also correlated with thenumber and percentages of eosinophils present in BALFas well as with total protein.3 This might also suggest thatother eosinophil granule proteins or plasma proteins con-tribute to alterations of surfactant function after local aller-gen challenge. In line with the explanation that plasmaproteins are responsible for surfactant dysfunction,6,24 wefound that BALF albumin showed a strong correlationwith surfactant dysfunction after allergen challenge. Albu-min as the most abundant plasma protein was present inBALF in higher amounts compared with ECP and fibrino-gen. In contrast to albumin, the plasma protein fibrinogendid not correlate to surfactant function, although it wasalso elevated after allergen challenge and was present inBALF in higher amounts than ECP. Thus, albumin but notfibrinogen appears to contribute to plasma protein-inducedsurfactant dysfunction in asthmatic subjects after segmen-tal allergen challenge.

In vitro experiments with purified ECP proved theconcept that this protein indeed induces alterations ofsurfactant function. The openness of surfactant-contain-ing capillaries that simulate airway architecture wasmarkedly reduced by ECP in a concentration-dependentmanner (Fig 3). In addition, the effect of ECP on surfac-tant function was reversed by anti-ECP. The in vitro con-centrations of ECP leading to significant alterations ofsurfactant function were 3 orders of magnitude higherthan ECP concentrations found in BALF from asthmaticsubjects in our clinical study. Although the exact ECPconcentration in the hypophase of the airway lining fluidis unknown, it is tempting to speculate that ECP levels in

the surfactant-containing hypophase are much higherthan BALF levels would suggest because the BAL pro-cedure with 100 mL saline dilutes BALF levels of ECP.Interestingly, looking at ECP levels in proximal and dis-tal airway compartments by performing sequentiallavage, Schmekel and Venge25 demonstrated that ECPconcentrations in the proximal airway compartment weresignificantly higher than in the distal compartment.Moreover, comparison of induced sputum and bronchialwashings as material from the proximal airway compart-ment with BAL as material of preferentially alveolar ori-gin provides further evidence for the differential distri-bution of ECP in the lungs of patients with mild asthma.Median levels of ECP in induced sputum were up to 770times higher than in BAL samples. In contrast, levels ofalbumin in induced sputum were only 4 times highercompared with BAL samples in these patients.26 Thesedata demonstrate that ECP concentrations in the proxi-mal airways of asthmatic patients reach levels that arecomparable with our in vitro assay. Therefore, weassume that ECP might be sufficient to induce a surfac-tant dysfunction in this airway compartment.

The inhibitory potential of ECP was compared withtwo well-known inhibitors of surfactant function albuminand fibrinogen.24 All three proteins induced functionalsurfactant alterations in a concentration-dependent fash-ion. However, the mass concentration that induce half-maximal inhibition of surfactant function was different.ECP was more potent in disturbing surfactant functionthan fibrinogen followed by albumin. This finding under-lines the potential relevance of ECP for the inhibition ofsurfactant function and adds a novel protein inhibitor ofsurfactant function to our current knowledge.

The mechanisms by which ECP interferes with sur-factant might be different compared with albumin or fi-brinogen. On the one hand, it has been shown that ECPis capable of forming pores into lipid bilayers,27 suggest-

FIG 6. Volume densities of ULV of surfactant suspensions con-taining natural bovine surfactant (Alveofact) at a phospholipidconcentration of 0.1 mg/mL alone (control) and after addition ofalbumin, fibrinogen, or ECP. Proteins were used at concentrationsof 100 µg/mL. Values are given as mean ± SEM. *P < .05 comparedwith control.

502 Hohlfeld et al J ALLERGY CLIN IMMUNOL

MARCH 2004

Mechanism

s of allergy

ing an enzymatic activity acting on phospholipids. Onthe other hand, we have demonstrated in this study thatECP induces structural changes of surfactant vesicles dif-ferent from those induced by albumin or fibrinogen.Although albumin showed little changes and fibrinogenonly induced moderate alterations to the structure ofmultilamellar surfactant vesicles, ECP markedly convert-ed those multilamellar bodies into unilamellar vesicles atsimilar protein concentrations. In addition, the effect ofECP on surfactant function was blocked by anti-ECP,suggesting a specific action of ECP on phospholipidsrather than the interaction of the total protein with phos-pholipid layer formation.

Although we have clearly demonstrated that ECPalters pulmonary surfactant function and structure, itremains unresolved whether ECP induces airwayobstruction mainly through a surfactant-dependentmechanism. In a recent study, it has been shown thatinstillation of ECP into rat lungs consistently inducedtachypnea and reduction of tidal volumes.28 These ECP-induced respiratory responses were completely prevent-ed by perineural capsaicin treatment, suggestinginvolvement of c-fiber afferents rather than a surfactant-mediated mechanism. However, lung function variablesthat allow estimation of the degree of airway obstructionhave not been determined in this study. It may very wellbe that an obstructive airway function remained that wasnot detectable by solely analyzing the breathing pattern.To our knowledge, there is no study that has measuredpulmonary function variables in response to ECP thatanswers this question.

In conclusion, we present data in this study thatdemonstrate that ECP induces inhibition of surfactantfunction. These functional changes are accompanied bypronounced structural alterations of surfactant inducedby ECP, a protein that is more potent than albumin or fi-brinogen. It is tempting to speculate that in patients withasthma and an eosinophilic airway inflammation, ECPreleased from eosinophils has an impact on airway cal-iber through a surfactant-dependent mechanism.

The skillful technical assistance of Marion Schaël and MartinaHanke is greatly acknowledged.

REFERENCES

1. Enhorning G. Pulmonary surfactant function in alveoli and conductingairways. Can Respir J 1996;3:21-7.

2. Hohlfeld J, Fabel H, Hamm H. The role of pulmonary surfactant inobstructive airways disease. Eur Respir J 1997;10:482-91.

3. Hohlfeld JM, Ahlf K, Enhorning G, Balke K, Erpenbeck VJ, PetschalliesJ, et al. Dysfunction of pulmonary surfactant in asthmatics after segmen-tal allergen challenge. Am J Respir Crit Care Med 1999;159:1803-9.

4. Jarjour NN, Enhorning G. Antigen-induced airway inflammation inatopic subjects generates dysfunction of pulmonary surfactant. Am JRespir Crit Care Med 1999;160:336-41.

5. Kurashima K, Fujimura M, Matsuda T, Kobayashi T. Surface activity ofsputum from acute asthmatic patients. Am J Respir Crit Care Med1997;155:1254-9.

6. Seeger W, Grube C, Günther A, Schmidt R. Surfactant inhibition by plas-ma proteins: differential sensitivity of various surfactant preparations.Eur Respir J 1993;6:971-7.

7. Heeley EL, Hohlfeld JM, Krug N, Postle AD. Phospholipid molecularspecies of bronchoalveolar lavage fluid after local allergen challenge inasthma. Am J Physiol (Lung Cell Mol Physiol) 2000;278:L305-11.

8. Venge P. The human eosinophil in inflammation. Agents Action1990;29:122-6.

9. Lacoste JY, Bousquet J, Chanez P, Van Vyve T, Simony-Lafontaine J,Lequeu N, et al. Eosinophilic and neutrophilic inflammation in asthma,chronic bronchitis, and chronic obstructive pulmonary disease. J AllergyClin Immunol 1993;92:537-48.

10. Fujimoto K, Kubo K, Matsuzawa Y, Sekiguchi M. Eosinophil cationicprotein levels in induced sputum correlate with the severity of bronchialasthma. Chest 1997;112:1241-7.

11. National Heart, Lung, and Blood Institute. International consensus reporton diagnosis and treatment of asthma. Eur Respir J 1992;5:601-41.

12. Cockcroft DW, Killian DN, Mellon JJ, Hargreave FE. Bronchoreactivityto inhaled histamine: a method and clinical survey. Clin Allergy1977;7:235-43.

13. Krug N, Erpenbeck VJ, Balke K, Petschallies J, Tschernig T, HohlfeldJM, et al. Cytokine profile of bronchoalveolar lavage-derived CD4+,CD8+, and γd T cells in people with asthma after segmental allergenchallenge. Am J Respir Cell Mol Biol 2001;25:125-31.

14. Lowry OH, Rosebrough NJ, Farr AL, Randall RJ. Protein measurementwith the folin phenol reagent. J Biol Chem 1951;193:265-75.

15. Liu M, Wang L, Li E, Enhorning G. Pulmonary surfactant will secure freeairflow through a narrow tube. J Appl Physiol 1991;71:742-8.

16. Enhorning G, Holm BA. Disruption of pulmonary surfactant’s ability tomaintain openness of a narrow tube. J Appl Physiol 1993;74:2922-7.

17. Weibel ER. A simplified morphometric method for estimating diffusingcapacity in normal and emphysematous human lungs. Am Rev Respir Dis1973;107:579-88.

18. Rosner BA. Fundamentals in Biostatistics. 4th ed. Belmont (CA):Duxberry Press; 1995.

19. Enhorning G, Duffy LC, Welliver RC. Pulmonary surfactant maintainspatency of conducting airways in the rat. Am J Respir Crit Care Med1995;151:554-6.

20. Hohlfeld JM. The role of pulmonary surfactant in asthma. Respir Res2002;3:4.

21. Woolley KL, Adelroth E, Woolley MJ, Ellis R, Jordana M, O’Byrne PM.Effects of allergen challenge on eosinophils, eosinophil cationic protein,and granulocyte-macrophage colony-stimulating factor in mild asthma.Am J Respir Crit Care Med 1995;151:1915-24.

22. Shaver JR, Zangrilli JG, Cho SK, Cirelli RA, Pollice M, Hastie AT, et al.Kinetics of the development and recovery of the lung from IgE-mediatedinflammation: dissociation of pulmonary eosinophilia, lung injury, andeosinophil-active cytokines. Am J Respir Crit Care Med 1997;155:442-8.

23. Peebles RSJ, Liu MC, Adkinson NFJ, Lichtenstein LM, Hamilton RG.Ragweed-specific antibodies in bronchoalveolar lavage fluids and serumbefore and after segmental lung challenge: IgE and IgA associated witheosinophil degranulation. J Allergy Clin Immunol 1998;101:265-73.

24. Seeger W, Stohr G, Wolf HRD, Neuhof H. Alteration of surfactant func-tion due to protein leakage: special interaction with fibrin monomer. JAppl Physiol 1985;58:326-38.

25. Schmekel B, Venge P. The distribution of myeloperoxidase, eosinophilcationic protein, albumin and urea in sequential bronchoalveolar lavage.Eur Respir J 1991;4:517-23.

26. Fahy JV, Wong H, Liu J, Boushey HA. Comparison of samples collectedby sputum induction and bronchoscopy from asthmatic and healthy sub-jects. Am J Respir Crit Care Med 1995;152:53-8.

27. Young JDE, Peterson CGB, Venge P, Cohn ZA. Mechanism of membranedamage mediated by human eosinophil cationic protein. Nature1986;321:613-6.

28. Lee LY, Gu Q, Gleich GJ. Effects of human eosinophil granule-derivedcationic protein on C-fiber afferents in the rat lung. J Appl Physiol2001;91:1318-26.