Nerve growth factor levels and localisation in human asthmatic bronchi

Subjects

This study included 14 subjects, of whom seven were patients with clinically stable, mild-to-moderate bronchial asthma (as defined in the International Consensus Report on Diagnosis and Management of Asthma, 1992 29), and seven were included as control subjects. The asthmatic patients were recruited from the authors' outpatient clinic. The clinical characteristics of the included subjects are displayed in tables 1 and 2⇓⇓. Each subject underwent a medical history, physical examination, and a pulmonary function testing by spirometry (forced expiratory volume in one second (FEV₁)) (Autospiro, Minato AS 500; Mediprom, Paris, France).

Selection criteria for the asthmatic patients included a well-documented history of asthma and bronchial hyperresponsiveness, as assessed by a methacholine challenge test (18–2,340 µg) at a screening visit, according to a standardised procedure, as described previously 30.

In the group of asthmatics (three males and four females, aged 21–26 yrs, median age 24 yrs), all but one were atopic, as indicated by weal and flare responses to one or more common allergen extracts used in the skin-prick testing. Fifteen common allergen extracts (Stallergènes, Fresnes, France) were tested according to a standard protocol. One patient informed that he had a food allergy to coriander, which provoked asthma symptoms, which was not included in the skin-prick test. The nonallergic asthmatic patient had a well-defined history of exercise-induced asthma. The asthmatic subjects had a mean±sem baseline FEV₁ prior to bronchoscopy of 84.4±5.5% (range 61.0–102.9) of the predicted normal values, and their geometric mean provocation dose causing a 20% fall in FEV₁ (PD₂₀) was 255.7 µg (table 1⇑).

Control subjects (three males and four females, aged 21–38 yrs, median age 25 yrs) were all healthy nonatopic, nonallergic volunteers. The nonasthmatic subjects had a mean±sem baseline FEV₁ prior to bronchoscopy of 100.8±2.7% (range 90.3–113.1) of predicted normal values, and they exhibited no airway hyperresponsiveness, with a PD₂₀ >2,340 µg methacholine in all subjects (table 2⇑).

All bronchodilators were withheld for at least 8 h before bronchoscopy. Most subjects were nonsmokers, except two subjects in each group (tables 1 and 2⇑⇑). The study was conducted out of the pollen season, and none of the allergic subjects had any animals or were exposed to the allergens against which they were sensitised for at least 1 month prior to the study. In addition, all subjects were free of respiratory infection for at least 1 month before the study, as well as from regular medication, other than the occasional use of inhaled bronchodilators in the group of asthmatics, or contraceptives. No systemic or inhaled steroids were used by any subject included in the study. The study was approved by the ethical committee of the authors' hospital, and all patients gave their written informed consent.

Fibreoptic bronchoscopy

Patients underwent fibroscopy between 4 days–5 weeks after the methacholine challenge test. After premedication with diazepam (10 mg per os), local anaesthesia with 1% adrenaline-free xylocaine was directly applied to the upper respiratory tract, while a flexible fibreoptic bronchial fibroscope (BF P30; Olympus, Paris, France) was inserted through the nose into the trachea. Bronchial biopsies were taken from the subsegmental division of the main bronchi in the left lower and upper lobes using alligator forceps (Radial Jaw; Microvasive, La Garenne-Colombes, France). Bronchoalveolar lavage (BAL) was performed in the subsegment in the right middle lobe, by instilling 50 mL of sterile isotonic saline four times, at 37°C, gently aspirated immediately after instillation, as recommended by the National Institutes of Health Workshop Summary and Guidelines, 1991 31. Patients received oxygen during bronchoscopy, an intravenous access was prepared, and adrenaline was readily available. Nebulisation with salbutamol was performed before and after the fibroscopy. Oximetry was recorded throughout the procedure and for 4 h after fibroscopy, with spirometry performed every hour for 4 h.

Preparation of bronchial biopsies and bronchoalveolar lavage fluid samples

Bronchial biopsies were gently extracted from the forceps, and immediately placed in 10% buffered formaldehyde fixative at 4°C for 3 h. Subsequently, the biopsies were embedded in paraffin and kept until use for immunohistochemistry.

The recovered BAL fluid (BALF) was kept on ice and immediately transported to the laboratory. The fluid was strained through a single layer of sterile cotton gauze to remove mucus. Cells were separated from the fluid phase by centrifugation at 400×g for 10 min. An aliquot of cells resuspended in phosphate-buffered saline was taken to perform cell counting, to assess cell viability, and to characterise the different cell populations by cytocentrifugation, and May Giemsa Grünewald staining. Total cell counts and percentages of different cell types (differential cell counts) are presented in table 3⇓. The cell-free BALF was then centrifuged at 1500×g for 15 min at 4°C to remove debris, then aliquoted and stored at −80°C until it was analysed.

Determination of nerve growth factor by enzyme-linked immunosorbent assay

The levels of NGF protein were quantified in 1:2 diluted BALF from asthmatic patients and control subjects with a commercially available, highly sensitive NGF-specific two-site enzyme-linked immunosorbent assay (ELISA)-kit according to the procedure indicated by the manufacturer (Promega, Madison, WI, USA), and as reported by other authors 11. Briefly, 96-well immunoplates (MaxiSorp^TM; Nunc, Roskilde, Denmark) were coated with a polyclonal goat anti-human NGF antibody in a coating buffer (25 mM carbonate buffer, pH 9.7). After an overnight incubation at 4°C, plates were washed (20 mM Tris-HCl, 150 mM NaCl with 0.05% (v/v) Tween®-20), and incubated in a blocking buffer (supplied by the manufacturer) for 1 h. The diluted BAL samples and the standard recombinant human NGF were incubated in the wells at room temperature for at least 6 h, and washed. Both samples and the recombinant NGF were diluted in the blocking buffer, as suggested by the manufacturer. Rat monoclonal anti-NGF antibody (0.25 µg·mL⁻¹) was added for an overnight incubation at 4°C. After washing, anti-rat horseradish peroxidase-conjugated immunoglobulin (Ig)G was added for a 2.5 h incubation period. Finally, the substrate (0.02% 3,3′,5,5′-tetramethylbenzidine and 0.01% hydrogen peroxidase) was added. The colourimetric reaction was stopped after 10 min with 1 M phosphoric acid, and the optical density was measured at 450 nm. All measurements were performed in duplicate. In control experiments, the primary antibody was withheld, leading to the expected undetectable signal. The technique allowed detection of NGF in the range of 3.9–500 pg·mL⁻¹.

Nerve growth factor immunolocalisation on bronchial biopsy sections

Serial 4 µm sections were deparaffinised in xylene, stepwise rehydrated through graded ethanol, and washed in Tris-buffered saline (TBS; 20 mM Trizma base, 150 mM NaCl, pH 7.6). Following antigen retrieval treatment with 0.1% pronase E in TBS for 8 min at room temperature (Sigma Chemical Co., St. Louis, MO, USA), unspecific protein binding was blocked with a blocking buffer containing foetal bovine serum (FBS) (10% v/v) (GIBCO BRL, Cergy Pontoise, France), and bovine serum albumin (BSA) (1% w/v) (Sigma) in TBS for 30 min at room temperature. Sections were then incubated overnight at 4°C with a rabbit anti-human NGF polyclonal antibody (1:100) (Santa Cruz Biotechnology, Santa Cruz, CA, USA), followed by incubation for 2 h at room temperature with a donkey biotinylated anti-rabbit IgG antiserum (1:200) (Amersham Life Science, Orsay, France), an Extravidin-alkaline phosphatase conjugate (Sigma), and finally with Fast Red TR/Naphtol AS-MX (Sigma) for development of the immune reaction. The anti-human NGF antibody recognises an epitope mapping the amino-terminus of the mature form of NGF of human origin, noncrossreactive with brain-derived neurotrophic factor, neurotrophin-3 or -4. Negative controls were performed by replacing the primary antibody by its control isotype immunoglobulin from the same species, and was performed in parallel for the different samples. All antibodies were diluted in blocking buffer. Sections were washed extensively in TBS between incubations with antibodies, conjugate and substrate. Finally, sections were washed in distilled water and counterstained with Mayer's haematoxylin, briefly dried and mounted in a gelatine-glycerol solution. A qualitative evaluation of the sections were performed by light microscopy using a Zeiss Axiophot (Oberkochen, Germany) at a magnification of ×200, ×400 and/or ×640.

Quantification of nerve growth factor-immunopositive inflammatory cells infiltrating the bronchial tissue

Slides were coded and examined under a light microscope (magnification ×400) (Zeiss Axiophot) by three investigators. NGF-immunopositive infiltrating inflammatory cells were counted in the submucosa, excluding the epithelium. The images were then digitalised with a high-performance charge-coupled device (CCD) camera (Cohu Inc., San Diego, CA, USA), and areas were measured by an image analyser (SIS analysis; Olympus). NGF-immunopositive inflammatory cells were counted in 2–3 different sections from each specimen, and counts were averaged. Results are expressed as number of NGF-immunopositive inflammatory cells per mm².

Statistical analysis

Group data were expressed as mean values±sem and range. Differences between groups were evaluated with the nonparametric Mann-Whitney sum-rank test. Values of p<0.05 were considered significant.

Nerve growth factor protein levels in bronchoalveolar lavage fluid

Detectable levels of NGF were determined in BALF from control subjects (53.9±20.2 pg·mL⁻¹). Significantly higher (121.5%, p<0.05) levels of NGF protein were observed in BALF from the asthmatic patients (119.4±30.6 pg·mL⁻¹) as compared with control subjects (fig. 1⇓).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.—

Nerve growth factor (NGF) protein levels in bronchoalveolar fluid from control subjects and asthmatic patients. NGF levels were quantified by enzy-melinked immunosorbent assay. Individual values are represented as dots and bars indicate mean±sem*. : p<0.05 between groups.

Nerve growth factor immunolocalisation in bronchial sections

NGF was immunolocalised on bronchial biopsy sections from control subjects and asthmatic patients. An intense NGF-immunolabelling was observed within the bronchial epithelium (fig. 2⇓ a–d), even though epithelium from the asthmatic bronchi showed mild shedding (fig. 2c and d⇓). The bronchial smooth muscle layer was also intensely labelled (fig. 2b and d⇓), whereas less intense immunoreactivity was observed in the connective tissue (fibroblasts) (fig. 2⇓). Furthermore, an intense positive signal for NGF was observed within infiltrating inflammatory cells in the submucosa (fig. 2c–e⇓). The majority of these cells had a mononuclear appearance. The negative control (fig. 2f⇓), performed by replacing the primary antibody by its control isotype from the same species, was entirely negative.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.—

Nerve growth factor (NGF) immunolocalisation on bronchial biopsy sections from control subjects (a, b) and asthmatic patients (c–f). Intense labelling was observed in bronchial epithelium (Ep), smooth muscle cells (Smc) and infiltrating inflammatory cells (Ic) in the submucosa, showing cytoplasmic localisation, both in control subjects and asthmatic patients. A negative control is shown (f), where the anti-human NGF antibody was replaced by its control isotype immunoglobulin.

Quantification of infiltrating nerve growth factor-immunopositive cells in the bronchial submucosa

In bronchial biopsies from asthmatics, a significant increase in the number of NGF-immunopositive cells infiltrating the submucosa was determined in comparison with the submucosa of control subjects (30.4±7.3 and 12.2±3.4, respectively, 148.6%, p<0.05).