Study population

We performed a prospective longitudinal study including bird researchers and/or managers at the Urban Pest Surveillance and Control Service (UPSCS) of the Barcelona Public Health Agency (ASPB), Municipal Institute of Parks and Gardens in Barcelona and employees of private urban pest control firms (n = 101). These individuals underwent a medical interview regarding their professional and domestic exposure to birds and fungi and pulmonary function tests to rule out respiratory symptoms. Blood extraction was performed in all subjects and after blood centrifugation (3000 rpm for 10 min) serum samples were obtained and stored at − 80 °C until analysis.

Patients with HP due to pigeon exposure (n = 5) who were diagnosed, according to the criteria proposed by Schuyler and Cornier [5], were included in the study as the control group for western blot analyses. The diagnosis was made when at least 4 major and 2 minor criteria were present and if other conditions with similar characteristics were ruled out. The major criteria were symptoms consistent with HP, evidence of appropriate antigen exposure in medical history and/or detection of specific precipitins in serum and/or bronchoalveolar lavage (BAL), findings consistent with HP on chest plain films or chest computed tomography, lymphocytosis in BAL (when performed), histological changes consistent with HP, and/or positive SIC. The minor criteria were bilateral basal crackles, decrease in DLCO, and/or arterial hypoxemia, at least after exercise.

Antigen extract preparation

Blood from different types of birds (parakeet, small parrot and parrot) was collected and centrifuged to obtain serum samples. Commercially available pigeon serum (Rockland Immunochemicals Inc., Limerick, Ireland), Aspergillus fumigatus, Penicillium frequentans and Mucor mucedo extracts (Bial-Aristegui, Bilbao, Spain) were used. Protein concentration of avian serums was determined by the Bicinchoninic acid (BCA) method (Pierce Chemical Co., Rockford, USA).

Specific IgG antibody determination

Specific IgG antibodies against avian and fungal proteins were determined in serum samples of all individuals of the study using a direct ELISA method.

Wells of high-binding microtiter plates (Costar, Cambridge, USA) were coated with avian serums and fungal extracts (2 µg protein/well) in coupling buffer (Na2CO3/NaHCO3, 0.2 M, pH 9.6) overnight at 4 °C. The wells were blocked (300 µL/well) with PBS/1% BSA for 1 h at 37 °C. After washing three times with PBST (PBS with 0.1% Tween-20), plates were incubated for 1 h at 37 °C with 200 µL/well of serum samples and negative and positive controls at 1/500 dilution in PBS/0.1% Gelatin/0.02% Tween-20. After three washes with PBST, plates were incubated for 1 h at 37 °C with 200 µL/well of mouse anti-human IgG-HRP antibody (SouthernBiotech, Birmingham, USA) at 1:1000 dilution in PBS/0.1% Gelatin/0.02% Tween-20. Wells were washed three times and the reaction was developed with 4.9% H2O2 and 3,3′,5,5′-tetramethylbenzidine in acetate buffer (pH 5.5) at 25 °C in the dark. After 20 min, the reaction was stopped with 2 M H2SO4 and read at 450 nm with a plate reader (Infinite M Nano, Tecan, Männedorf, Switzerland). Results were expressed as optical density (OD) and Magellan software (Tecan Austria GmbH, Grödig, Austria) was used for calculating the results.

Values above the mean plus 2 SDs of the results obtained in a control population of 30 healthy individuals previously studied in our laboratory were considered positive. The cut-offs were 0.284, 0.193, 0.294, 0.348, 0.508, 0.417, 0.258 and 0.687 OD450nm for pigeon, small parrot, parrot, parakeet, duck, Aspergillus fumigatus, Mucor mucedo and Penicillium frequentans, respectively.

Pulmonary function testing

A complete pulmonary function study including spirometry, static pulmonary volumes and diffusion capacity of the lung for carbon monoxide (DLCO) was carried out in all participants on a MasterLab system (MasterLab, Jaeger, Germany), in accordance with the guidelines of the European Respiratory Society and American Thoracic Society [16]. Static pulmonary volumes were measured using the plethysmography method and the diffusion study was carried out using the single-breath carbon monoxide method. The theoretical values were the ones proposed by the European Respiratory Society and American Thoracic Society [17, 18]. A restrictive ventilatory pattern was defined as forced vital capacity (FVC) < 85% of the predicted value with forced expiratory volume in 1 s (FEV1)/FVC ratio > 80% in the absence of a static lung volume study, together with a total lung capacity (TLC) < 80% of the predicted value, when these tests were performed. An obstructive ventilatory pattern was established on the basis of FEV1/FVC ratio < 70% together with FEV1 < 80% of the predicted value. The DLCO was considered decreased at values of < 80% of the predicted value.

1D/2D electrophoresis gel and western blot analysis

For 1D electrophoresis performance, sample buffer was previously prepared diluting 2-mercaptoethanol at 1/10 in Laemmli buffer 4x (Bio-Rad, Madrid, Spain). Avian serums (850 µg protein/1D gel for pigeon and small parrot serums) were then mixed with sample buffer at a 3:1 proportion. After heating the samples for 5 min at 90 °C in the water bath, samples were loaded onto preparative Criterion TGX Stain-Free gels 4–20% (Bio-Rad, Madrid, Spain) in accordance with the manufacturer’s manual. For 2D electrophoresis, avian serums (100 µg protein/strip) mixed with rehydration solution (pure H2O/8.5 M Urea/2% Chaps/0.5% IPG buffer/0.002% Bromophenol blue) were loaded onto immobilized pH gradient strips (IPG strip, pH 3–10, 7 cm, General Electric Healthcare, Boston, USA). Isoelectric focusing (IEF) was then carried out using the IPGphor IEF system (General Electric Healthcare, Boston, USA) and following the manufacturer’s instructions. IPG strips were then equilibrated and loaded onto Mini-PROTEAN TGX Stain-Free gels 4–20% (7 cm, IPG/prep, Bio-Rad, Madrid, Spain). After 1D or 2D electrophoresis, proteins were stained with Coomassie blue or blotted to a PVDF membrane using the Trans-Blot Turbo transfer system (Bio-Rad, Madrid, Spain). After blocking the membranes with nonfat dry milk (Bio-Rad, Madrid, Spain) for 1 h at RT and gentle shaking, washing steps were carried out with PBST (PBS with 0.05% Tween-20).

Western blots were performed by incubating serum samples diluted 1/100 in PBS for 2 h at RT with gentle shaking. After washing three times with PBST, membranes were incubated with goat anti-human IgG-HRP (SouthernBiotech, Birmingham, USA) diluted 1/2000 (for 1D blots) or 1/4000 (for 2D blots) in PBS for 1 h at RT. Membranes were then washed three times with PBST and incubated with the enzyme substrate for 5 min at RT (Kit Clarity Western ECL Substrate, Bio-Rad, Madrid, Spain), in accordance with the supplier’s recommendations. Image acquisition was carried out with the Odyssey Fc Imaging System and the Image Studio Lite software (LI-COR, Nebraska, USA). Spot analyses in 2D blots were performed using the PDQuest 2D software (Bio-Rad, Madrid, Spain). Healthy volunteers (HV; n = 3) without previous avian exposure and patients with BRHP due to pigeons (n = 5), who were diagnosed between 2009 and 2016 at our center (Hospital Universitari Vall d’Hebron, Barcelona, Spain) according to the criteria proposed by Schuyler and Cornier [5] were included as controls.

Mass spectrometry analysis

Spots recognized by HP patients but not by workers nor HV were excised from 2D stained gels and digested with trypsin according to standard protocols. The tryptic peptides were analyzed with a maXis UHR-Qq-TOF mass spectrometer (Bruker Daltonics, Bremen, Germany) and subsequently sequenced using both the DeNovo and the Spider Homology Search tools from the PEAKS Studio 5.3 software (Bioinformatics Solutions Inc., Waterloo, ON, Canada).

Statistical analysis

The normal distribution of the data was evaluated with the Shapiro–Wilk test. For categorical variables, data were analysed using chi-square or Fisher’s exact test and are expressed as absolute numbers and their corresponding percentages. For quantitative variables, data were analysed using Mann–Whitney U-test or Unpaired T-test, as appropriate, and data are shown as medians and ranges. Analyses were conducted using GraphPad Prism 6 for Windows (version 6.01, GraphPad Software Inc, San Diego, California, USA) and IBM SPSS Statistics (version 26, IBM Corporation, Armonk, New York, USA). Differences with a p-value < 0.05 (two-tailed) were considered to be significant.