Galectin-3: an early predictive biomarker of modulation of airway remodeling in patients with severe asthma treated with omalizumab for 36 months

Eight patients were treated with the monoclonal antibody anti-IgE omalizumab for 36 months. The patients’ clinical features (Table 1) were described in a previous paper, where we showed that the original RBM thickness and eosinophil infiltration were reduced in a substantial proportion of severe asthmatics after 1 year of treatment with omalizumab, thus emphasizing the possible role of this agent in airway remodelling in severe persistent allergic asthma [6].

The study was approved by the Ethics Committee of Orlandi General Hospital of Bussolengo, Verona, Italy.

As previously reported before treatment and at 12 months [6], each patient underwent bronchoscopy 36 months after starting anti-IgE treatment. Bronchial biopsy specimens were obtained using a flexible bronchoscope (Pentax FB19-TX, Langley, UK). Specimens were collected from the right middle lobe, fixed in formalin for 10 h at 4 °C, and embedded in paraffin. The blocks were cut into 3-µm sections using a rotary microtome.

The bronchial sections were mounted onto glass slides, dewaxed, rehydrated, and stained with hematoxylin-eosin (Hematoxylin and Eosin Stain, Carl Roth GmbH + Co. KG, Germany).

The area of the RBM was measured using computer-aided digital morphometry with a DFC 320 Leica color digital camera attached to a Leica Microsystems DMLA light microscope. Digital images of the biopsies were captured at high power using a 100× lens and analyzed using QWin software (Leica Microsystems). Bronchial RBM thickness was measured according to the recommendations of the American Thoracic Society/European Respiratory Society using the orthogonal intercept method. A 100 × 100-μm grid was randomly overlaid on hematoxylin-eosin-stained sections. Orthogonal intercepts were measured from the intersections of the grid with the RBM-epithelium junction to the RBM-subepithelial area junction. At least 40 measures were obtained every 20 μm. Morphometric analysis was performed by two blinded operators. The arithmetic means of the intercept measures were calculated using the formula τ = π/4 × arithmetic mean of orthogonal intercepts [8].

Bronchial biopsies collected 36 months after starting treatment with omalizumab were considered suitable for examination if at least 1.0 mm of RBM length and 0.1 mm² of subepithelial area were morphometrically preserved. Eosinophils were quantified in the area extending 50 μm under the RBM and expressed as number of eosinophils/mm² of subepithelium (mean of two samples per patient). The eosinophil count was performed by two blinded operators [9].

In order to obtain complete data on the bronchial inflammatory pattern, we evaluated neutrophilic infiltration before treatment and after 36 months. This variable was not investigated in our previous paper [6]. Biopsies were considered suitable for examination when there was at least 1.0 mm of RBM length and 0.1 mm² of subepithelial area. Neutrophils were quantified in the area extending 50 μm under the RBM and expressed as number of neutrophils/mm² of subepithelium. The neutrophil count was performed by two blinded operators [9].

The dewaxed tissues (3-µm section from each sample) were dried in a vacuum centrifuge and resuspended in 0.1 M ammonium bicarbonate, pH 7.9. Proteins were extracted from tissue as previously described [7]. Briefly, tissue was homogenized in solution buffer (0.1 M ammonium bicarbonate, pH 7.9) treated with RapiGest SF reagent (Waters Corporation, Milford, MA, USA) and incubated with stirring, first at 100 °C for 20 min and then at 80 °C for 2 h. Subsequently, the protein concentration was assayed using the SPN^TM Protein Assay kit (G-Biosciences, Maryland Heights, MO, USA), and 5 ± 0.5 µg of protein from each sample was digested with trypsin (Sequencing Grade Modified Trypsin, Promega, Madison, WI, USA) using a 1:50 (w/w) enzyme/substrate ratio at 37 °C overnight. The next morning, an additional aliquot of enzyme was added at an enzyme/substrate ratio of 1:100 (w/w), and digestion was continued for 4 h. Enzyme digestion was stopped by addition of trifluoroacetic acid to reach a pH of 2, and the digested samples were desalted and enriched using PepClean columns (Pierce Biotechnology, Rockford, IL, USA).

The resulting peptide mixtures were analyzed using multidimensional protein identification technology (MudPIT) [10] based on two-dimensional chromatography coupled to tandem mass spectrometry (2DC-MS/MS). Peptides were identified to correlate the experimental tandem mass spectra with the theoretical peptide sequences obtained by the in silico digestion of a human protein database (approximately 230,000 entries) downloaded from the NCBI website (www.​ncbi.​nlm.​nih.​gov).

Mass spectra data were processed using Bioworks version 3.3.1 based on the SEQUEST algorithm (University of Washington, licensed to Thermo Finnigan Corp., San José, CA, USA) and the following parameters: Xcorr scores greater than 1.5 for singly charged peptide ions and 2.0 and 2.5 for doubly and triply charged ions, respectively; peptide probability ≤0.001; and protein consensus score value ≥10. These filters guaranteed that the resulting proteins had a p value of ≤0.001. The false-positive peptide ratio, which was calculated through the reverse database, was less than 3%.

The statistical analysis was performed using R software. Biological and technical replicates were evaluated by hierarchical clustering [11] using an in-house R script based on the XlsReadWrite, clue, and clValid libraries (http://​cran.​r-project.​org). The Euclidean distance metric was applied and an agglomerative coefficient was calculated.

We measured five parameters (eosinophils, smooth muscle proteins, periostin, keratins, and RBM) in eight patients (OR, n = 4; NOR, n = 4) and separated them according to time (T0, and T36) to evaluate significant differences between OR and NOR. For each group, significant differences were also evaluated between baseline (T0) and T36.

As for eosinophils (cells/mm²), the t test (p ≤ 0.05) was used to evaluate the difference between the patient groups (OR and NOR) and treatment time (T0 and T36). The same comparisons were performed for periostin, Gal-3, smooth muscle proteins, and keratins. In this case, the average SEQUEST score values were evaluated using the Wilcoxon test because proteomic data are not normally distributed (Shapiro–Wilk test).

Variation in the proteins analyzed was also evaluated by calculating the fold change, as previously reported [12]; this was defined as the natural logarithm of the ratio T36/T0. In addition, the protein lists were analyzed and aligned using multidimensional algorithm protein map software (MAProMa), which provides quantitative indexes based on the SEQUEST score, such as DAVE and DCI [13]. DAVE is an index of the ratio between the two conditions compared and, therefore, indicates different amounts of each protein under the two different conditions. When the protein is not present in the reference (baseline), DAVE is +2.0, whereas if the protein is not present in the sample (in our case after treatment, T36), DAVE is −2.0. Consequently, intermediate DAVE values indicate different amounts of protein under the two conditions compared. DAVE values >|0.4| indicate significant variations. Finally, the t test was used to evaluate significant differences in the average DAVE and fold change values.