Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

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

Galectin-3: an early predictive biomarker of modulation of airway remodeling in patients with... Background: Bronchial asthma is a heterogeneous disease characterized by three cardinal features: chronic inflam- mation, variable airflow obstruction, and airway hyperresponsiveness. Asthma has traditionally been defined using nonspecific clinical and physiologic variables that encompass multiple phenotypes and are treated with nonspecific anti-inflammatory therapies. Based on the modulation of airway remodeling after 12 months of anti-immunoglobulin E (IgE) treatment, we identified two phenotypes (omalizumab responder, OR; and non-omalizumab responder, NOR) and performed morphometric analysis of bronchial biopsy specimens. We also found that these two phenotypes were correlated with the presence/absence of galectin-3 (Gal-3) at baseline (i.e., before treatment). The aims of the present study were to investigate the histological and molecular effects of long-term treatment (36 months) with anti-IgE and to analyze the behavior of OR and NOR patients. Methods: All patients were treated with the monoclonal antibody anti-IgE omalizumab for 36 months. The bronchial biopsy specimens were evaluated using morphometric, eosinophilic, and proteomic analysis (MudPIT ). New data were compared with previous data, and unsupervised cluster analysis of protein profiles was performed. Results: After 36 months of treatment with omalizumab, reduction of reticular basement membrane (RBM) thickness was confirmed in OR patients (Gal-3-positive at baseline); similarly, the protein profiles (over 500 proteins identified) revealed that, in the OR group, levels of proteins specifically related to fibrosis and inflammation (e.g., smooth muscle and extracellular matrix proteins (including periostin), Gal-3, and keratins decreased by between 5- and 50-fold. Eosin- ophil levels were consistent with molecular data and decreased by about tenfold less in ORs and increased by twofold to tenfold more in NORs. This tendency was confirmed (p < 0.05) based on both fold change and DAVE algorithms, thus indicating a clear response to anti-IgE treatment in Gal-3-positive patients. Conclusions: Our results showed that omalizumab can be considered a disease-modifying treatment in OR. The pro- teomic signatures confirmed the presence of Gal-3 at baseline to be a biomarker of long-term reduction in bronchial RBM thickness, eosinophilic inflammation, and muscular and fibrotic components in omalizumab-treated patients with severe asthma. Our findings suggest a possible relationship between Gal-3 positivity and improved pulmonary function. *Correspondence: canonica@unige.it Anna Maria Riccio, Pierluigi Mauri contributed equally to this manuscript. Department of Biomedical Sciences, Personalized Medicine Clinic Asthma and Allergy, Humanitas University, Rozzano, Milan, Italy Full list of author information is available at the end of the article © The Author(s) 2017. This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/ publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Riccio et al. Clin Transl Allergy (2017) 7:6 Page 2 of 10 Keywords: Anti-IgE, Omalizumab, Severe asthma, Galectin-3, Biomarker, Airway remodeling, Bronchial biopsy, Proteomics, Eosinophils Background another or not at all. The reasons for these differential Bronchial asthma is a heterogeneous disease character- responses are still unknown [5]. ized by three cardinal features: chronic inflammation, We performed morphometric analysis on bronchial variable airflow obstruction, and airway hyperrespon - biopsy specimens before and after anti-IgE treatment siveness [1]. The heterogeneity of asthma extends beyond to investigate modulation (or not) of airway remodeling clinical symptoms, response to therapy, age at onset, after 12 months of treatment with omalizumab. We iden- duration of disease process, extent of bronchial narrow- tified two phenotypes of severe asthma: the omalizumab ing, sensitivity to triggering agents, airway inflammatory responder phenotype (OR) and the non-omalizumab pattern, and the immune response. Accordingly, clas- responder phenotype (NOR) [6]. Proteomic analysis sifying asthma into several smaller, more homogeneous of the specimens showed that these patient subgroups subgroups, the so-called phenotypes, makes it easier to were characterized by different levels of galectin-3 (Gal- define the underlying mechanisms of the disease, admin - 3) in bronchial tissue at baseline and after 12  months of ister more effective mechanism-based therapy, and treatment with omalizumab [7]. In the present study, improve prediction of disease course. we extended our morphometric, eosinophilic, and pro- Technological advances during the last decade have teomic analyses by investigating bronchial biopsy speci- facilitated clinical and molecular research, thus making it mens collected from the same patients after 36  months possible to combine phenotyping approaches to optimize of treatment with omalizumab and compared molecular asthma management and personalized medicine [2]. data with data on bronchial reticular basement mem- The US Severe Asthma Research Program (SARP) brane (RBM) thickness and bronchial eosinophilic and recently analyzed an extensive dataset of patients with neutrophilic infiltration. severe and nonsevere asthma in order to identify and The aims of the present study were to evaluate prot - describe robust subgroups of asthma patients with spe- eomic signatures and modifications in RBM thickness cific features [3] that could guide personalized therapy, with respect to long-term anti-IgE treatment and to although this approach has not yet been verified. Using investigate the behaviour of OR and NOR. a variety of statistical analyses, the authors interrogated the dataset to identify the characteristics that most accu- Methods rately distinguish between subgroups within the study Patients population. This approach is known as clustering [1]. Eight patients were treated with the monoclonal anti- Asthma has traditionally been defined using nonspe - body anti-IgE omalizumab for 36  months. The patients’ cific clinical and physiological variables that encompass clinical features (Table  1) were described in a previous multiple phenotypes and has been treated using nonspe- paper, where we showed that the original RBM thickness cific anti-inflammatory therapies. Recent molecular and and eosinophil infiltration were reduced in a substantial genetic studies have identified clinical and inflammatory proportion of severe asthmatics after 1 year of treatment phenotypes associated with specific biomarkers. with omalizumab, thus emphasizing the possible role Biomarkers for inflammation driven by type-2 helper T of this agent in airway remodelling in severe persistent lymphocytes (Th2), including elevated fractional exhaled allergic asthma [6]. nitric oxide (FeNO) levels, blood/sputum eosinophil counts, and serum periostin levels, have helped to iden- Table 1 Patient characteristics tify a Th2-high molecular phenotype of asthma. Treat - Patients (n) 8 ment of Th2-high asthmatic patients with biologic agents Age (years) 47.0 ± 9.7 targeting immunoglobulin E (IgE) and the canonical Sex ratio (M/F) 5/3 Th2-related cytokines interleukin (IL) 4, IL-5, and IL-13 Smokers/nonsmokers 1/7 is proving to be efficacious [4]. However, we do not yet Body mass index 23.8 ± 3.1 know the biomarkers able to identify a Th2-low asthma Total plasma IgE (kU/L) 309.4 ± 218.2 phenotype and potentially guide therapy. Although tar- Forced expiratory volume in the first second 52% pred ± 14% geted biologic agents are generally efficacious in treat - Asthma control test score 11.3 ± 2.8 ing various phenotypes of asthma and allergic disease, some patients respond better to one biologic agent than Data are expressed as mean ± SEM Riccio et al. Clin Transl Allergy (2017) 7:6 Page 3 of 10 The study was approved by the Ethics Committee of considered suitable for examination when there was at Orlandi General Hospital of Bussolengo, Verona, Italy. least 1.0  mm of RBM length and 0.1  mm of subepithe- lial area. Neutrophils were quantified in the area extend - Bronchial biopsy: collection and processing ing 50  μm under the RBM and expressed as number of As previously reported before treatment and at neutrophils/mm of subepithelium. The neutrophil count 12  months [6], each patient underwent bronchoscopy was performed by two blinded operators [9]. 36  months after starting anti-IgE treatment. Bronchial biopsy specimens were obtained using a flexible bron - Proteomic analysis choscope (Pentax FB19-TX, Langley, UK). Specimens The dewaxed tissues (3-µm section from each sample) were collected from the right middle lobe, fixed in forma - were dried in a vacuum centrifuge and resuspended in lin for 10 h at 4 °C, and embedded in paraffin. The blocks 0.1  M ammonium bicarbonate, pH 7.9. Proteins were were cut into 3-µm sections using a rotary microtome. extracted from tissue as previously described [7]. Briefly, tissue was homogenized in solution buffer (0.1 M ammo - Morphometric analysis nium bicarbonate, pH 7.9) treated with RapiGest SF The bronchial sections were mounted onto glass slides, reagent (Waters Corporation, Milford, MA, USA) and dewaxed, rehydrated, and stained with hematoxylin-eosin incubated with stirring, first at 100  °C for 20  min and (Hematoxylin and Eosin Stain, Carl Roth GmbH  +  Co. then at 80  °C for 2  h. Subsequently, the protein con- TM KG, Germany). centration was assayed using the SPN Protein Assay The area of the RBM was measured using computer- kit (G-Biosciences, Maryland Heights, MO, USA), and aided digital morphometry with a DFC 320 Leica color 5 ± 0.5 µg of protein from each sample was digested with digital camera attached to a Leica Microsystems DMLA trypsin (Sequencing Grade Modified Trypsin, Promega, light microscope. Digital images of the biopsies were Madison, WI, USA) using a 1:50 (w/w) enzyme/substrate captured at high power using a 100× lens and analyzed ratio at 37 °C overnight. The next morning, an additional using QWin software (Leica Microsystems). Bronchial aliquot of enzyme was added at an enzyme/substrate RBM thickness was measured according to the recom- ratio of 1:100 (w/w), and digestion was continued for 4 h. mendations of the American Thoracic Society/Euro - Enzyme digestion was stopped by addition of trifluoro - pean Respiratory Society using the orthogonal intercept acetic acid to reach a pH of 2, and the digested samples method. A 100 × 100-μm grid was randomly overlaid on were desalted and enriched using PepClean columns hematoxylin-eosin-stained sections. Orthogonal inter- (Pierce Biotechnology, Rockford, IL, USA). cepts were measured from the intersections of the grid The resulting peptide mixtures were analyzed using with the RBM-epithelium junction to the RBM-subepi- multidimensional protein identification technology thelial area junction. At least 40 measures were obtained (MudPIT) [10] based on two-dimensional chromatog- every 20  μm. Morphometric analysis was performed raphy coupled to tandem mass spectrometry (2DC-MS/ by two blinded operators. The arithmetic means of the MS). Peptides were identified to correlate the experi - intercept measures were calculated using the formula mental tandem mass spectra with the theoretical peptide τ = π/4 × arithmetic mean of orthogonal intercepts [8]. sequences obtained by the in silico digestion of a human protein database (approximately 230,000 entries) down- Evaluation of the eosinophilic infiltrate loaded from the NCBI website (www.ncbi.nlm.nih.gov). Bronchial biopsies collected 36  months after start- ing treatment with omalizumab were considered suit- Statistical and clustering analyses able for examination if at least 1.0  mm of RBM length Mass spectra data were processed using Bioworks ver- and 0.1  mm of subepithelial area were morphometri- sion 3.3.1 based on the SEQUEST algorithm (Univer- cally preserved. Eosinophils were quantified in the sity of Washington, licensed to Thermo Finnigan Corp., area extending 50  μm under the RBM and expressed as San José, CA, USA) and the following parameters: number of eosinophils/mm of subepithelium (mean of Xcorr scores greater than 1.5 for singly charged peptide two samples per patient). The eosinophil count was per - ions and 2.0 and 2.5 for doubly and triply charged ions, formed by two blinded operators [9]. respectively; peptide probability ≤0.001; and protein consensus score value ≥10. These filters guaranteed that Evaluation of the neutrophilic infiltrate the resulting proteins had a p value of ≤0.001. The false- In order to obtain complete data on the bronchial inflam - positive peptide ratio, which was calculated through the matory pattern, we evaluated neutrophilic infiltration reverse database, was less than 3%. before treatment and after 36  months. This variable was The statistical analysis was performed using R soft - not investigated in our previous paper [6]. Biopsies were ware. Biological and technical replicates were evaluated Riccio et al. Clin Transl Allergy (2017) 7:6 Page 4 of 10 by hierarchical clustering [11] using an in-house R script Analysis of bronchial tissue in the OR patients revealed based on the XlsReadWrite, clue, and clValid libraries a consistent reduction in RBM thickness; conversely, (http://cran.r-project.org). The Euclidean distance met - analysis of the NOR patients revealed an increase in RBM ric was applied and an agglomerative coefficient was thickness despite treatment with omalizumab (Fig. 1). calculated. As for subepithelial eosinophils, our data showed We measured five parameters (eosinophils, smooth that the eosinophilic infiltrate doubled (7–15/mm ) in muscle proteins, periostin, keratins, and RBM) in eight NOR patients after 36 months. Analysis of biopsy speci- patients (OR, n  =  4; NOR, n  =  4) and separated them mens from the OR patients, on the other hand, revealed according to time (T0, and T36) to evaluate significant the presence of eosinophils to be about 10 times lower differences between OR and NOR. For each group, sig - (55–5/mm ) (Fig.  2). The difference between OR and nificant differences were also evaluated between baseline NOR was significant (p < 0.05; fold change  = ln[T36/T0]) (T0) and T36. (Additional file 1). As for eosinophils (cells/mm ), the t test (p  ≤  0.05) Evaluation of the neutrophilic infiltrate revealed no dif - was used to evaluate the difference between the patient ferences between the two groups. Moreover, the number groups (OR and NOR) and treatment time (T0 and T36). of neutrophils in bronchial specimens at baseline, after The same comparisons were performed for periostin, both 12 months and 36 months was very low and negligi- Gal-3, smooth muscle proteins, and keratins. In this case, ble (data not shown). the average SEQUEST score values were evaluated using the Wilcoxon test because proteomic data are not nor- Protein profile mally distributed (Shapiro–Wilk test). MudPIT made it possible to identify 546 distinct proteins Variation in the proteins analyzed was also evaluated (see Additional file  2 for a complete list of the proteins by calculating the fold change, as previously reported identified for each sample). Using the MAProMa software, [12]; this was defined as the natural logarithm of the each protein list was automatically plotted onto a 2D map ratio T36/T0. In addition, the protein lists were analyzed according to the theoretical molecular weight (MW) and and aligned using multidimensional algorithm protein isoelectric point (pI) of the proteins identified. A repre - map software (MAProMa), which provides quantitative sentative example is reported in Additional file  3, which indexes based on the SEQUEST score, such as DAVE and shows the 2D maps corresponding to all the proteins iden- DCI [13]. DAVE is an index of the ratio between the two tified, thus confirming the possibility of characterizing conditions compared and, therefore, indicates different proteins with extreme theoretical MW values (<10  kDa) amounts of each protein under the two different condi - and proteins with extreme pI values (<4 or >10). tions. When the protein is not present in the reference In our previous study [6], the main differences at (baseline), DAVE is +2.0, whereas if the protein is not 12  months affected the cytoskeleton, mainly smooth present in the sample (in our case after treatment, T36), muscle and keratins, and ECM proteins. With respect DAVE is −2.0. Consequently, intermediate DAVE val- to ECM proteins, analysis of the biopsy specimens at ues indicate different amounts of protein under the two 36  months revealed the same trend of proteomic signa- conditions compared. DAVE values >|0.4| indicate sig- ture for patients who were Gal-3-positive before treat- nificant variations. Finally, the t test was used to evalu - ment (OR); in contrast, protein behavior was different ate significant differences in the average DAVE and fold in patients who were Gal-3-negative before treatment change values. (NOR). In the OR group, smooth muscle proteins (including Results myosins, tropomyosins, and actins) had decreased to very We evaluated the effect of long-term anti-IgE treatment low levels at 36 months (from 265 to 8 [aggregate score]; (36  months) on RBM thickness, eosinophilic and neu- average ln[T36/T0], about –4.0; DAVE, –1.89) in the OR trophilic infiltrates, and proteomic profiles by analyzing group (Gal-3-positive biopsies). In the NOR group (Gal- bronchial biopsy specimens from patients with severe 3-negative biopsies), protein levels increased from 5 to 74 asthma (Table 1). (average ln[T36/T0], around +2.5; DAVE, +1.75) (Fig. 3; Additional files 1, 4). The difference between the two Histological analysis groups was almost significant (p < 0.01). All bronchial biopsies were evaluated using morphomet- As for keratins, the trend at 36  months of treatment ric analysis of RBM thickness. was more consistent for OR (Gal-3 positive) (about six- Differences were observed in the behavior of the two fold less, from 1800 to 300 as the average score; average groups (OR and NOR) between baseline and 36  months ln[T36/T0], −2; DAVE −1.45) than NOR (both DAVE of anti-IgE treatment. and average ln[T36/T0], low stringent). In this case the Riccio et al. Clin Transl Allergy (2017) 7:6 Page 5 of 10 Bronchial reticular basement membrane (RBM) thickness changes NOR OR BASELINE AFTER 12M AFTER 36M Fig. 1 Changes in bronchial RBM thickness in omalizumab responders (OR) and non-omalizumab responders (NOR) at baseline and after 12 and 36 months of anti-IgE treatment the decreasing trend at 36 months. Specifically, periostin was reduced in all OR patients (from 117 to 0 average score; average ln[T36/T0], −4.6; DAVE −2.0), whereas in the NOR group, it was either stable (NOR3) or increased (NOR1 and NOR4), except for NOR2, in whom the periostin level decreased. Figure  5 shows the behav- ior (ln[T36/T0]) of periostin in OR and NOR over time (p < 0.05; see also Additional files 1, 4). Gal-3 was stable at 12 months, but was absent in OR at 36  months (Gal-3-positive at T0). Gal-3 was not detect- able at 36 months in NOR (Gal-3-negative at T0) (Addi- tional file 5). As reported above, NOR patients, who were char- acterized as Gal-3-negative at T0 (before treatment), presented the same behavior after 36  months of treat- Fig. 2 Evaluation of bronchial eosinophilic infiltration (cells/mm ) in ment; however, the trend for one patient (NOR2) after omalizumab responders (OR) and non-omalizumab responders (NOR) 36  months of treatment was more similar to that of OR using the natural logarithm of the score fold change, (ln[ T36/T0]). Negative values indicate a decrease in the eosinophil count at T36; than NOR patients for each of the main parameters mon- positive values indicate an increase in the eosinophil count at T36. itored. In fact, in NOR2, both smooth muscle proteins OR and NOR are classified according to whether they experienced and keratins had decreased or were absent at 36 months, a reduction in RBM thickness (OR) or no reduction in RBM thickness as was the case for OR patients. Similarly, eosinophils (NOR) after 12 months of anti-IgE treatment and periostin were, respectively, reduced and absent in NOR2, as in OR and in contrast to the remaining NOR. Table 2 summarizes the levels of the main protein classes difference was not significant (p  = 0.19; see Fig. 4; Addi- and eosinophils before treatment (T0) and 36 months after tional files 1, 4). treatment for each OR and NOR sample. Of note, both the ECM proteins, which were present almost exclusively at ln[T36/T0] (Additional file  1) and DAVE index (Additional baseline in OR (Gal-3-positive) patients, also maintained Bronchial RBM thickness (µm) Riccio et al. Clin Transl Allergy (2017) 7:6 Page 6 of 10 Clustering analysis Protein lists from the 36-month biopsy specimens were compared and used to perform an unsupervised clus- tering analysis. Thus, OR (Gal-3-positive), who were correctly segregated at T0 (see Fig.  6a), were again con- firmed to be part of the same group at 36  months. It is noteworthy that patient NOR2 was in this group. In addi- tion, the remaining three NOR patients (NOR1, NOR3, and NOR4, all Gal-3-negative) were segregated in a dif- ferent cluster (see Fig.  6b), thus confirming them to be part of the same NOR group again at 36 months of treat- ment. These findings confirm the different behavior at 36 months of patient NOR2, who was segregated with the OR group (see Table 2). Fig. 3 Changes in abundance levels, expressed as the natural loga- rithm of the score fold change (ln[ T36/T0]) for smooth muscle pro- Discussion teins (SMPs) in omalizumab responders (OR) and non-omalizumab Biomarkers of severe asthma are currently a major responders (NOR) at baseline ( T0) and after 36 months ( T36) of treat- research area [14]. Nonetheless, available biomarkers [15] ment. A negative value indicates a decrease in SMPs at T36; a positive are still far from enabling us to select appropriate candi- value indicates an increase in SMPs at T36. OR and NOR are classified dates for biologics [16, 17]. The need for molecular analy - according to whether they experienced a reduction in RBM thickness (OR) or no reduction in RBM thickness (NOR) after 12 months of anti- sis to define phenotypes has been discussed [18], and IgE treatment omics sciences have been proposed as a useful approach [19]. In this context, current advances in proteomics, mainly those based on mass spectrometry, facilitate the discov- ery-driven studies of new biomarkers in respiratory dis- eases and improve the clinical reliability of biomarkers for asthma [20]. The main focus of the present study was the extension of the histological and proteomic analyses to 36  months in a group of patients with severe asthma treated with omalizumab that we had previously studied at baseline and at 12  months [6]. In the previous study, we did not have control biopsy specimens from untreated severe asthmatic patients and/or healthy persons, since bron- choscopy and bronchial biopsy are invasive procedures that are not readily applicable in these groups. Although the number of samples was not large for the usual statistical analysis, the data obtained illustrate a Fig. 4 Changes in abundance levels, expressed as the natural loga- specific trend for both groups of patients (OR and NOR). rithm of the score fold change (ln[ T36/T0]) for keratins in omalizumab In particular, our results showed that specimens from OR responders (OR) and non-omalizumab responders (NOR) at baseline patients (Gal-3-positive at baseline) confirm the reduc - ( T0) and after 36 months ( T36). A negative value indicates a decrease tion in the thickness of bronchial RBM after 36  months in keratins at T36; a positive value indicates an increase in keratins at of treatment with anti-IgE, thus emphasizing the possi- T36. In this case, the p value is high; however, if NOR2 is excluded, the p value falls to 0.002 (Additional file 1). OR and NOR are classified ble role of omalizumab in reducing airway remodeling according to whether they experienced a reduction in RBM thickness in some cases of persistent allergic asthma [6]. As for (OR) or no reduction in RBM thickness (NOR) after 12 months of anti- molecular components, proteomic analysis showed that IgE treatment mainly ECM, keratin, and smooth muscle proteins were further reduced at 36  months of therapy in Gal-3-posi- tive biopsies. file  6) confirm the different trends for OR and NOR in the Although Gal-3 was absent in OR at 36  months, this parameters examined at baseline and T36. They also high - does not put OR in the same situation as NOR; in fact, light the different behavior of NOR2 at T36. OR presented a different proteomic profile, consistent Riccio et al. Clin Transl Allergy (2017) 7:6 Page 7 of 10 low serum periostin. This predictive role for periostin was not confirmed in recent clinical trials [22]. In the present study, periostin was not consistently present in the bron- chial biopsies of either group, and levels decreased fur- ther only in OR after 36 months of treatment. Therefore, based on our data, periostin cannot be regarded as a pre- dictive biomarker of modulation of airway remodeling. With regard to the bronchial inflammatory pat - tern, high eosinophil counts in the bronchi have been observed in patients who are eligible for anti-IgE [23] and can be considered one of the biomarkers of the Th2 asthma phenotype. Omalizumab is effective in the treat - ment of eosinophilic airway inflammation in patients with allergic asthma [24]. In the present study, a further Fig. 5 Changes in abundance levels expressed as the natural loga- decrease (compared with 12  months) in eosinophils rithm of the score fold change (ln[ T36/T0]) for periostin in omali- in the biopsy specimens was detected in patients who zumab responders (OR) and non-omalizumab responders (NOR) at were Gal-3-positive at baseline (OR), whereas eosino- baseline ( T0) and after 36 months ( T36) of anti-IgE treatment. A nega- phil counts increased in Gal-3-negative patients (NOR). tive value indicates a decrease in periostin at T36; a positive value These findings are consistent with recent data reporting indicates an increase in periostin at T36. OR and NOR are classified according to whether they experienced a reduction in RBM thickness increased and decreased Gal-3 levels in eosinophilic and (OR) or no reduction in RBM thickness (NOR) after 12 months of anti- neutrophilic asthma, respectively [25]. Gal-3 might also IgE treatment predict decreased eosinophilic inflammation in the bron - chi in patients receiving omalizumab. Finally, the further decrease in smooth muscle proteins with the reduction in RBM thickness and eosino- (i.e., actins and myosins) after 36 months is interesting and phil counts. In this context, the lack of Gal-3 in OR at thus supports the hypothesis of improved pulmonary func- 36 months, combined with the decrease in protein levels tion observed in some omalizumab-treated asthmatics [26]. reported above, may be correlated with an improvement If this was the case, Gal-3 would be a very useful tool for in remodelling and inflammatory state. predicting clinical improvement in pulmonary function. With respect to periostin, Hanania et al. [21] found that This observation is of particular clinical interest, since the patients with high serum periostin had fewer exacerba above-mentioned studies showed that, as far as FEV is tions after treatment with omalizumab than patients with Table 2 Levels of  eosinophils (cells/mm ) and  abundance (score) for  smooth muscle proteins, periostin, and  keratins at baseline (T0) and after 36 months of anti-IgE treatment (T36) OR1 OR2 OR3 OR4 NOR1 NOR2 NOR3 NOR4 Eosinophils T0 19 31 58 115 0 10 10 9 T36 3 9 4 5 11 3 20 27 Smooth muscle proteins T0 288 80 198 494 0 0 20 0 T36 10 20 0 0 168 0 30 98 Periostin T0 30 160 130 150 0 30 120 0 T36 0 0 0 0 20 0 106 20 Keratins T0 1990 1902 1630 2047 1404 1910 1892 1676 T36 574 276 100 266 2106 40 1330 1400 Galectin-3 T0 10 10 30 20 0 0 0 0 T36 0 0 0 0 0 0 0 0 Increased values (or substantially unchanged) at T36 compared to T0 are reported in italic. For more details about available proteins see Additional file 2 Riccio et al. Clin Transl Allergy (2017) 7:6 Page 8 of 10 a Baseline b After 36-months Fig. 6 Unsupervised hierarchical clustering of protein lists obtained by analyzing lung biopsies from omalizumab responders (OR, blue) and non- omalizumab responders (NOR, red). a Before (baseline). b After 36 months of anti-IgE treatment. The clusters obtained were based on SEQUEST score value and proved to be well structured (agglomerative coefficients >0.7). OR and NOR are classified according to whether they experienced a reduction in RBM thickness (OR) or no reduction in RBM thickness (NOR) after 12 months of anti-IgE treatment concerned, not all patients improved with omalizumab. In PROXIMA involves an ancillary study to explore protein responsive patients, omalizumab can be considered a dis- biomarkers in urine and plasma and to characterize them ease-modifying treatment and Gal-3 a potential biomarker. according to severe allergic asthma and treatment effects. Our findings confirm the role of Gal-3 as a biomarker In this context, the possible correlation between Gal-3 of long-term reduction in bronchial RBM, eosinophil and evolution of pulmonary function with omalizumab inflammation, and muscular components in omali - will be investigated in greater depth. zumab-treated severe asthmatics. In addition, our data support the molecular model, indicating that the pres- Conclusions ence of Gal-3 may have a role in the dissociation of the We investigated the long-term effects of anti-IgE treat - IgE–FcεRI complex and enhance the effectiveness of anti- ment. In particular, proteomic signatures combined with IgE therapy for remodelling in Gal-3-positive patients histological data enabled us to confirm the following: [7]. Further studies based on larger samples or a cellular model are required to confirm the role of Gal-3 in the • In bronchial biopsy samples from severe asthmatics, dissociation of the IgE–FcεRI complex. Gal-3 acts as a biomarker of modulation of airway Bronchial analysis is useful for investigating the origin remodeling upon treatment with omalizumab. and mechanisms of disease; however, in clinical practice, • There seems to be a relationship between Gal-3 posi - it is necessary to analyze less invasive samples, such as tivity and improvement in pulmonary function. urine and blood. We intend to perform such an analysis • In severe asthmatics treated with omalizumab, Gal-3 in the context of the PROXIMA study (Patient-Reported is a sufficiently effective biomarker of both short- and Outcomes and Xolair In the Management of Asthma), long-term reduction in the thickening of bronchial an observational, multicenter, cross-sectional, prospec- RBM, eosinophilic inflammation, and fibrotic and tive cohort study conducted at 25 centers [27]. In fact, muscular protein components. OR OR1 1 OR1 OR1 OR2 OR OR4 4 OR4 OR OR41 OR OR2 2 OR2 OR3 OR2 OR3 OR3 OR4 NOR2 NOR3 NOR3 NOR3 NOR1 NOR1 NOR1 OR3 NOR1 NOR1 NOR2 NOR2 NOR2 NOR4 NOR4 NOR4 NOR3 Riccio et al. Clin Transl Allergy (2017) 7:6 Page 9 of 10 study, data analysis and interpretation, and drafting of the manuscript. DFL Additional files performed the experiments, analyzed the data, and helped to draft the manu- script. RR performed the experiments, analyzed the data, and interpreted Additional file 1. Natural logarithm (ln) of fold change between baseline the results. DD analyzed the data and interpreted the results. BL performed ( T0) and long term anti-IgE treatment ( T36) for each subject. Significant the experiments. DRW recruited the patients and collected clinical data. MC values (ln[Fold Change]>|0.6| are are reported in red or blue: positive performed bronchoscopy and collected bronchial biopsy samples. CA contrib- (blue) and negative (red) values indicate increase and decrease at T36, uted to data analysis and participated in drafting the manuscript and in the respectively. Specifically, for ln(Fold Change) >0.6 (blue) increase at T36; critical review thereof. CGW was the lead investigator and was responsible for on the contrary, if it is <-0.6 (red) decrease at T36. SMPs: smooth muscle designing the study and drafting the manuscript. Author information: RAM proteins. Actual Fold Change was 0 (0/n, or 0/0, see Table 2); for avoiding and MP contributed equally to this work. All authors read and approved the invalid logarithm it was put at 1/n or 1/1; Actual Fold Change ∞ (n/0); for final manuscript. avoiding invalid logarithm it was put at n/1. * p-value; T-test2 is without NOR2 subject, because at T36 its behaviour is similar to ORs (se Fig. 6); in Author details bold significant T-tests; ** #/mm2 ratio; *** score ratio. Respiratory Diseases and Allergy Unit, IRCCS AOU San Martino-IST, Univer- sity of Genoa, Genoa, Italy. Proteomics and Metabolomics Unit, Institute Additional file 2. Complete list of identified proteins in OR (Responders) for Biomedical Technologies, CNR, Milan, Italy. National Centre for Respira- and NOR (Non-Responders) bronchial biopsies collected after 36 months tory Pharmacoeconomics and Pharmacoepidemiology, CESFAR, Verona, Italy. of anti-IgE treatment. Accession number NCBI, Uniprot entry, Protein 4 5 Respiratory Unit, Mater Salutis Hospital, Legnago, Verona, Italy. Department name, pI and MW, Spectral Count (SpC) and Score were reported for each of Biomedical Sciences, Personalized Medicine Clinic Asthma and Allergy, protein. The symbol (*) was used to indicate the smooth muscle proteins. Humanitas University, Rozzano, Milan, Italy. The symbol (†) was used to indicate the periostin. The symbol (°) was used to indicate the keratins. Acknowledgements Additional file 3. Virtual 2D map in logarithmic scale of FFPE bronchial This work was partially supported by ARMIA (Associazione Ricerca Malattie biopsies, generated using MAProMa software (546 proteins). Proteins are Immunologiche ed Allergiche) and the Italian CNR FLAGSHIPs Projects, funded plotted according to their theoretical pI and MW. A color/shape code by the MIUR, “InterOmics”. The authors thank Danila Vella for statistical support. is assigned to each protein according to SEQUEST score value. Proteins with score ≤15 are reported as yellow/triangle, proteins with score ≥35 Competing interests are reported as red/circle, and proteins in the range 15–35 are reported as The authors declare that they have no competing interests. blue/square. The protein lists are reported in Additional file 2. Availability of data and materials Additional file 4. Changes of abundance levels, calculated by DAVE The datasets supporting the conclusions of this article are included within the algorithm from MAPROMA software [9], for eosinoplis, smooth muscle article and its additional files. proteins, periostin, keratins and Gal-3 in OR and NOR patients at baseline ( T0) and after 36 months ( T36) of anti-IgE treatment. Negative value indi- Ethics approval cates decrease at T36; on the contrary, positive value indicates increase at This study was approved by the local Ethics Committee. T36. OR and NOR classification is related to reduction (OR) or not (NOR) of RBM thickness after 12 months of anti-IgE treatment. Funding Additional file 5. Changes of abundance levels, expressed as natural The authors declare that the research was not funded. logarithm of score fold change (ln[ T36/T0]), for Gal-3 in OR and NOR patients at baseline ( T0) and after 36 months ( T36) of anti-IgE treatment. Received: 13 September 2016 Accepted: 22 February 2017 Negative value indicates decrease of Gal-3 at T36; on the contrary, positive value indicates increase of Gal-3 at T36 (see Additional file 1). OR and NOR classification is related to reduction (OR) or not (NOR) of RBM thickness after 12 months of anti-IgE treatment. Additional file 6. Differential Analysis (DAVE index) between baseline References ( T0) and long term anti-IgE treatment ( T36) for each subject, using DAVE 1. Brasier AR, Ju H. Analysis and predictive modeling of asthma phenotypes. index from MAPROMA software [9]. Significant values (DAVE > |0.4|) are Adv Exp Med Biol. 2014;795:273–88. reported in red or blue: positive (blue) and negative (red) values indicate 2. Wenzel SE. Asthma phenotypes: the evolution from clinical to molecular increase and decrease at T36, respectively. SMPs: smooth muscle proteins. approaches. Nat Med. 2012;18:716–25. Of note, DAVE algorithm, tpical of proteomics evaluation, was also 3. Jarjour NN, Erzurum SC, Bleecker ER, Calhoun WJ, Castro M, Comhair SA, applied to eosinophils; the obtained values resulted in good agreement et al. Severe asthma: lessons learned from the National Heart, Lung, and with evaluation obtained by ln[ T36/T0] (see also Additional file 1 and Blood Institute Severe Asthma Research Program. Am J Respir Crit Care compare Fig. 2 and Additional file 4). * p-value; T-test2 is without NOR2 Med. 2012;185:356–62. subject, because at T36 its behaviour is similar to ORs (see Fig. 6); in bold 4. Gandhi NA, Bennett BL, Graham NM, Pirozzi G, Stahl N, Yancopoulos GD. significant T-tests. Targeting key proximal drivers of type 2 inflammation in disease. Nat Rev Drug Discov. 2016;15(1):35–50. 5. Fajt ML, Wenzel SE. Asthma phenotypes and the use of biologic medica- tions in asthma and allergic disease: the next steps toward personalized Abbreviations care. J Allergy Clin Immunol. 2015;135:299–310. OR: omalizumab responder; NOR: non-omalizumab responder; IgE: immuno- 6. Riccio AM, Dal Negro RW, Micheletto C, De Ferrari L, Folli C, Chiappori A, globulin E; MudPIT: multidimensional proteomic identification technology; et al. Omalizumab modulates bronchial reticular basement membrane Gal-3: galectin-3; ECM: extracellular matrix; Th2: T-helper type-2 lymphocytes; thickness and eosinophil infiltration in severe persistent allergic asthma FeNO: fractional exhaled nitric oxide; IL: interleukin; pI: isoelectric point; RBM: patients. Int J Immunopathol Pharmacol. 2012;25:475–84. reticular basement membrane; FEV : forced expiratory volume in the first 7. Mauri P, Riccio AM, Rossi R, Di Silvestre D, Benazzi L, De Ferrari L, et al. second; MW: molecular weight; ln: natural logarithm; FC: fold change (ratio Proteomics of bronchial biopsies: galectin-3 as a predictive biomarker of T36/T0); P: p value. airway remodelling modulation in omalizumab-treated severe asthma patients. Immunol Lett. 2014;162(1 Pt A):2–10. Authors’ contributions 8. Hsia CC, Hyde DM, Ochs M, Weibel ER. ATS/ERS Joint Task Force on Quan- RAM contributed to the design of the study, data acquisition, the statistical titative Assessment of Lung Structure. An official research policy state - analysis, and drafting of the manuscript. MP contributed to design of the ment of the American Thoracic Society/European Respiratory Society: Riccio et al. Clin Transl Allergy (2017) 7:6 Page 10 of 10 standards for quantitative assessment of lung structure. Am J Respir Crit 20. Rossi R, De Palma A, Benazzi L, Riccio AM, Canonica GW, Mauri P. Care Med. 2010;181:394–418. Biomarker discovery in asthma and COPD by proteomic approaches. 9. Chetta A, Zanini A, Foresi A, Del Donno M, Castagnaro A, D’Ippolito R, Proteomics Clin Appl. 2014;8:901–15. et al. Vascular component of airway remodeling in asthma is reduced by 21. Hanania NA, Wenzel S, Rosen K, Hsieh HJ, Mosesova S, Choy DF, et al. high dose of fluticasone. Am J Respir Crit Care Med. 2003;167:751–7. Exploring the effects of omalizumab in allergic asthma: an analysis of bio - 10. Mauri P, Scigelova M. Multidimensional protein identification technology markers in the EXTRA study. Am J Respir Crit Care Med. 2013;187:804–11. for clinical proteomic analysis. Clin Chem Lab Med. 2009;47:636–46. 22. Hanania NA, Korenblat P, Chapman KR, Bateman ED, Kopecky P, Paggiaro 11. Zhao Y, Karypis G. Data clustering in life sciences. Mol Biotechnol. P, et al. Efficacy and safety of lebrikizumab in patients with uncontrolled 2005;31:55–80. asthma (LAVOLTA I and LAVOLTA II): replicate, phase 3, randomised, dou- 12. Vigani G, Di Silvestre D, Agresta AM, Donnini S, Mauri P, Gehl C et al. ble-blind, placebo-controlled trials. Lancet Respir Med. 2016;4:781–96. Molybdenum and iron mutually impact their homeostasis in cucumber 23. Busse W, Spector S, Rosén K, Wang Y, Alpan O. High eosinophil count: (Cucumis sativus) plants. New Phytol. 2017;213(3):1222–41. a potential biomarker for assessing successful omalizumab treatment 13. Mauri P, Dehò G. A proteomic approach to the analysis of RNA degrado- effects. J Allergy Clin Immunol. 2013;132:485–6. some composition in Escherichia coli. Methods Enzymol. 2008;447:99–117. 24. Djukanović R, Wilson SJ, Kraft M, Jarjour NN, Steel M, Chung KF. Eec ff ts 14. Arron JR, Choy DF, Scheerens H, Matthews JG. Noninvasive biomarkers of treatment with anti-immunoglobulin E antibody omalizumab on that predict treatment benefit from biologic therapies in asthma. Ann airway inflammation in allergic asthma. Am J Respir Crit Care Med. Am Thorac Soc. 2013;10(Suppl):S206–13. 2004;170:583–93. 15. Leung TF, Ko FW, Wong GW. Recent advances in asthma biomarker 25. Gao P, Gibson PG, Baines KJ, Yang IA, Upham JW, Reynolds PN, Hodge S, research. Ther Adv Respir Dis. 2013;7:297–308. James AL, Jenkins C, Peters MJ, Zhang J, Simpson JL. Anti-inflammatory 16. Chung KF. New treatments for severe treatment-resistant asthma: target- deficiencies in neutrophilic asthma: reduced galectin-3 and IL-1RA/IL-1β. ing the right patient. Lancet Respir Med. 2013;1:639–52. Respir Res. 2015;24(16):5. 17. De Ferrari L, Chiappori A, Bagnasco D, Riccio AM, Passalacqua G, Canonica 26. Baena-Cagnani CE, Teijeiro A, Canonica GW. Four-year follow-up in GW. Molecular phenotyping and biomarker development: Are we on our children with moderate/severe uncontrolled asthma after withdrawal way towards targeted therapy for severe asthma? Expert Rev Respir Med. of a 1-year omalizumab treatment. Curr Opin Allergy Clin Immunol. 2016;10:29–38. 2015;15:267–71. 18. Wenzel SE. Complex phenotypes in asthma: current definitions. Pulm 27. Chiappori A, De Ferrari L, Folli C, Mauri P, Riccio AM, Canonica GW. Pharmacol Ther. 2013;26:710–5. Biomarkers and severe asthma: a critical appraisal. Clin Mol Allergy. 19. Holgate ST. Trials and tribulations in identifying new biologic treatments 2015;13:20. for asthma. Trends Immunol. 2012;33:238–46. Submit your next manuscript to BioMed Central and we will help you at every step: • We accept pre-submission inquiries • Our selector tool helps you to find the most relevant journal • We provide round the clock customer support • Convenient online submission • Thorough peer review • Inclusion in PubMed and all major indexing services • Maximum visibility for your research Submit your manuscript at www.biomedcentral.com/submit http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Clinical and Translational Allergy Springer Journals

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

Loading next page...
 
/lp/springer-journals/galectin-3-an-early-predictive-biomarker-of-modulation-of-airway-Wji95O3sVb

References (30)

Publisher
Springer Journals
Copyright
Copyright © 2017 by The Author(s)
Subject
Medicine & Public Health; Allergology; Immunology; Pneumology/Respiratory System
eISSN
2045-7022
DOI
10.1186/s13601-017-0143-1
Publisher site
See Article on Publisher Site

Abstract

Background: Bronchial asthma is a heterogeneous disease characterized by three cardinal features: chronic inflam- mation, variable airflow obstruction, and airway hyperresponsiveness. Asthma has traditionally been defined using nonspecific clinical and physiologic variables that encompass multiple phenotypes and are treated with nonspecific anti-inflammatory therapies. Based on the modulation of airway remodeling after 12 months of anti-immunoglobulin E (IgE) treatment, we identified two phenotypes (omalizumab responder, OR; and non-omalizumab responder, NOR) and performed morphometric analysis of bronchial biopsy specimens. We also found that these two phenotypes were correlated with the presence/absence of galectin-3 (Gal-3) at baseline (i.e., before treatment). The aims of the present study were to investigate the histological and molecular effects of long-term treatment (36 months) with anti-IgE and to analyze the behavior of OR and NOR patients. Methods: All patients were treated with the monoclonal antibody anti-IgE omalizumab for 36 months. The bronchial biopsy specimens were evaluated using morphometric, eosinophilic, and proteomic analysis (MudPIT ). New data were compared with previous data, and unsupervised cluster analysis of protein profiles was performed. Results: After 36 months of treatment with omalizumab, reduction of reticular basement membrane (RBM) thickness was confirmed in OR patients (Gal-3-positive at baseline); similarly, the protein profiles (over 500 proteins identified) revealed that, in the OR group, levels of proteins specifically related to fibrosis and inflammation (e.g., smooth muscle and extracellular matrix proteins (including periostin), Gal-3, and keratins decreased by between 5- and 50-fold. Eosin- ophil levels were consistent with molecular data and decreased by about tenfold less in ORs and increased by twofold to tenfold more in NORs. This tendency was confirmed (p < 0.05) based on both fold change and DAVE algorithms, thus indicating a clear response to anti-IgE treatment in Gal-3-positive patients. Conclusions: Our results showed that omalizumab can be considered a disease-modifying treatment in OR. The pro- teomic signatures confirmed the presence of Gal-3 at baseline to be a biomarker of long-term reduction in bronchial RBM thickness, eosinophilic inflammation, and muscular and fibrotic components in omalizumab-treated patients with severe asthma. Our findings suggest a possible relationship between Gal-3 positivity and improved pulmonary function. *Correspondence: canonica@unige.it Anna Maria Riccio, Pierluigi Mauri contributed equally to this manuscript. Department of Biomedical Sciences, Personalized Medicine Clinic Asthma and Allergy, Humanitas University, Rozzano, Milan, Italy Full list of author information is available at the end of the article © The Author(s) 2017. This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/ publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Riccio et al. Clin Transl Allergy (2017) 7:6 Page 2 of 10 Keywords: Anti-IgE, Omalizumab, Severe asthma, Galectin-3, Biomarker, Airway remodeling, Bronchial biopsy, Proteomics, Eosinophils Background another or not at all. The reasons for these differential Bronchial asthma is a heterogeneous disease character- responses are still unknown [5]. ized by three cardinal features: chronic inflammation, We performed morphometric analysis on bronchial variable airflow obstruction, and airway hyperrespon - biopsy specimens before and after anti-IgE treatment siveness [1]. The heterogeneity of asthma extends beyond to investigate modulation (or not) of airway remodeling clinical symptoms, response to therapy, age at onset, after 12 months of treatment with omalizumab. We iden- duration of disease process, extent of bronchial narrow- tified two phenotypes of severe asthma: the omalizumab ing, sensitivity to triggering agents, airway inflammatory responder phenotype (OR) and the non-omalizumab pattern, and the immune response. Accordingly, clas- responder phenotype (NOR) [6]. Proteomic analysis sifying asthma into several smaller, more homogeneous of the specimens showed that these patient subgroups subgroups, the so-called phenotypes, makes it easier to were characterized by different levels of galectin-3 (Gal- define the underlying mechanisms of the disease, admin - 3) in bronchial tissue at baseline and after 12  months of ister more effective mechanism-based therapy, and treatment with omalizumab [7]. In the present study, improve prediction of disease course. we extended our morphometric, eosinophilic, and pro- Technological advances during the last decade have teomic analyses by investigating bronchial biopsy speci- facilitated clinical and molecular research, thus making it mens collected from the same patients after 36  months possible to combine phenotyping approaches to optimize of treatment with omalizumab and compared molecular asthma management and personalized medicine [2]. data with data on bronchial reticular basement mem- The US Severe Asthma Research Program (SARP) brane (RBM) thickness and bronchial eosinophilic and recently analyzed an extensive dataset of patients with neutrophilic infiltration. severe and nonsevere asthma in order to identify and The aims of the present study were to evaluate prot - describe robust subgroups of asthma patients with spe- eomic signatures and modifications in RBM thickness cific features [3] that could guide personalized therapy, with respect to long-term anti-IgE treatment and to although this approach has not yet been verified. Using investigate the behaviour of OR and NOR. a variety of statistical analyses, the authors interrogated the dataset to identify the characteristics that most accu- Methods rately distinguish between subgroups within the study Patients population. This approach is known as clustering [1]. Eight patients were treated with the monoclonal anti- Asthma has traditionally been defined using nonspe - body anti-IgE omalizumab for 36  months. The patients’ cific clinical and physiological variables that encompass clinical features (Table  1) were described in a previous multiple phenotypes and has been treated using nonspe- paper, where we showed that the original RBM thickness cific anti-inflammatory therapies. Recent molecular and and eosinophil infiltration were reduced in a substantial genetic studies have identified clinical and inflammatory proportion of severe asthmatics after 1 year of treatment phenotypes associated with specific biomarkers. with omalizumab, thus emphasizing the possible role Biomarkers for inflammation driven by type-2 helper T of this agent in airway remodelling in severe persistent lymphocytes (Th2), including elevated fractional exhaled allergic asthma [6]. nitric oxide (FeNO) levels, blood/sputum eosinophil counts, and serum periostin levels, have helped to iden- Table 1 Patient characteristics tify a Th2-high molecular phenotype of asthma. Treat - Patients (n) 8 ment of Th2-high asthmatic patients with biologic agents Age (years) 47.0 ± 9.7 targeting immunoglobulin E (IgE) and the canonical Sex ratio (M/F) 5/3 Th2-related cytokines interleukin (IL) 4, IL-5, and IL-13 Smokers/nonsmokers 1/7 is proving to be efficacious [4]. However, we do not yet Body mass index 23.8 ± 3.1 know the biomarkers able to identify a Th2-low asthma Total plasma IgE (kU/L) 309.4 ± 218.2 phenotype and potentially guide therapy. Although tar- Forced expiratory volume in the first second 52% pred ± 14% geted biologic agents are generally efficacious in treat - Asthma control test score 11.3 ± 2.8 ing various phenotypes of asthma and allergic disease, some patients respond better to one biologic agent than Data are expressed as mean ± SEM Riccio et al. Clin Transl Allergy (2017) 7:6 Page 3 of 10 The study was approved by the Ethics Committee of considered suitable for examination when there was at Orlandi General Hospital of Bussolengo, Verona, Italy. least 1.0  mm of RBM length and 0.1  mm of subepithe- lial area. Neutrophils were quantified in the area extend - Bronchial biopsy: collection and processing ing 50  μm under the RBM and expressed as number of As previously reported before treatment and at neutrophils/mm of subepithelium. The neutrophil count 12  months [6], each patient underwent bronchoscopy was performed by two blinded operators [9]. 36  months after starting anti-IgE treatment. Bronchial biopsy specimens were obtained using a flexible bron - Proteomic analysis choscope (Pentax FB19-TX, Langley, UK). Specimens The dewaxed tissues (3-µm section from each sample) were collected from the right middle lobe, fixed in forma - were dried in a vacuum centrifuge and resuspended in lin for 10 h at 4 °C, and embedded in paraffin. The blocks 0.1  M ammonium bicarbonate, pH 7.9. Proteins were were cut into 3-µm sections using a rotary microtome. extracted from tissue as previously described [7]. Briefly, tissue was homogenized in solution buffer (0.1 M ammo - Morphometric analysis nium bicarbonate, pH 7.9) treated with RapiGest SF The bronchial sections were mounted onto glass slides, reagent (Waters Corporation, Milford, MA, USA) and dewaxed, rehydrated, and stained with hematoxylin-eosin incubated with stirring, first at 100  °C for 20  min and (Hematoxylin and Eosin Stain, Carl Roth GmbH  +  Co. then at 80  °C for 2  h. Subsequently, the protein con- TM KG, Germany). centration was assayed using the SPN Protein Assay The area of the RBM was measured using computer- kit (G-Biosciences, Maryland Heights, MO, USA), and aided digital morphometry with a DFC 320 Leica color 5 ± 0.5 µg of protein from each sample was digested with digital camera attached to a Leica Microsystems DMLA trypsin (Sequencing Grade Modified Trypsin, Promega, light microscope. Digital images of the biopsies were Madison, WI, USA) using a 1:50 (w/w) enzyme/substrate captured at high power using a 100× lens and analyzed ratio at 37 °C overnight. The next morning, an additional using QWin software (Leica Microsystems). Bronchial aliquot of enzyme was added at an enzyme/substrate RBM thickness was measured according to the recom- ratio of 1:100 (w/w), and digestion was continued for 4 h. mendations of the American Thoracic Society/Euro - Enzyme digestion was stopped by addition of trifluoro - pean Respiratory Society using the orthogonal intercept acetic acid to reach a pH of 2, and the digested samples method. A 100 × 100-μm grid was randomly overlaid on were desalted and enriched using PepClean columns hematoxylin-eosin-stained sections. Orthogonal inter- (Pierce Biotechnology, Rockford, IL, USA). cepts were measured from the intersections of the grid The resulting peptide mixtures were analyzed using with the RBM-epithelium junction to the RBM-subepi- multidimensional protein identification technology thelial area junction. At least 40 measures were obtained (MudPIT) [10] based on two-dimensional chromatog- every 20  μm. Morphometric analysis was performed raphy coupled to tandem mass spectrometry (2DC-MS/ by two blinded operators. The arithmetic means of the MS). Peptides were identified to correlate the experi - intercept measures were calculated using the formula mental tandem mass spectra with the theoretical peptide τ = π/4 × arithmetic mean of orthogonal intercepts [8]. sequences obtained by the in silico digestion of a human protein database (approximately 230,000 entries) down- Evaluation of the eosinophilic infiltrate loaded from the NCBI website (www.ncbi.nlm.nih.gov). Bronchial biopsies collected 36  months after start- ing treatment with omalizumab were considered suit- Statistical and clustering analyses able for examination if at least 1.0  mm of RBM length Mass spectra data were processed using Bioworks ver- and 0.1  mm of subepithelial area were morphometri- sion 3.3.1 based on the SEQUEST algorithm (Univer- cally preserved. Eosinophils were quantified in the sity of Washington, licensed to Thermo Finnigan Corp., area extending 50  μm under the RBM and expressed as San José, CA, USA) and the following parameters: number of eosinophils/mm of subepithelium (mean of Xcorr scores greater than 1.5 for singly charged peptide two samples per patient). The eosinophil count was per - ions and 2.0 and 2.5 for doubly and triply charged ions, formed by two blinded operators [9]. respectively; peptide probability ≤0.001; and protein consensus score value ≥10. These filters guaranteed that Evaluation of the neutrophilic infiltrate the resulting proteins had a p value of ≤0.001. The false- In order to obtain complete data on the bronchial inflam - positive peptide ratio, which was calculated through the matory pattern, we evaluated neutrophilic infiltration reverse database, was less than 3%. before treatment and after 36  months. This variable was The statistical analysis was performed using R soft - not investigated in our previous paper [6]. Biopsies were ware. Biological and technical replicates were evaluated Riccio et al. Clin Transl Allergy (2017) 7:6 Page 4 of 10 by hierarchical clustering [11] using an in-house R script Analysis of bronchial tissue in the OR patients revealed based on the XlsReadWrite, clue, and clValid libraries a consistent reduction in RBM thickness; conversely, (http://cran.r-project.org). The Euclidean distance met - analysis of the NOR patients revealed an increase in RBM ric was applied and an agglomerative coefficient was thickness despite treatment with omalizumab (Fig. 1). calculated. As for subepithelial eosinophils, our data showed We measured five parameters (eosinophils, smooth that the eosinophilic infiltrate doubled (7–15/mm ) in muscle proteins, periostin, keratins, and RBM) in eight NOR patients after 36 months. Analysis of biopsy speci- patients (OR, n  =  4; NOR, n  =  4) and separated them mens from the OR patients, on the other hand, revealed according to time (T0, and T36) to evaluate significant the presence of eosinophils to be about 10 times lower differences between OR and NOR. For each group, sig - (55–5/mm ) (Fig.  2). The difference between OR and nificant differences were also evaluated between baseline NOR was significant (p < 0.05; fold change  = ln[T36/T0]) (T0) and T36. (Additional file 1). As for eosinophils (cells/mm ), the t test (p  ≤  0.05) Evaluation of the neutrophilic infiltrate revealed no dif - was used to evaluate the difference between the patient ferences between the two groups. Moreover, the number groups (OR and NOR) and treatment time (T0 and T36). of neutrophils in bronchial specimens at baseline, after The same comparisons were performed for periostin, both 12 months and 36 months was very low and negligi- Gal-3, smooth muscle proteins, and keratins. In this case, ble (data not shown). the average SEQUEST score values were evaluated using the Wilcoxon test because proteomic data are not nor- Protein profile mally distributed (Shapiro–Wilk test). MudPIT made it possible to identify 546 distinct proteins Variation in the proteins analyzed was also evaluated (see Additional file  2 for a complete list of the proteins by calculating the fold change, as previously reported identified for each sample). Using the MAProMa software, [12]; this was defined as the natural logarithm of the each protein list was automatically plotted onto a 2D map ratio T36/T0. In addition, the protein lists were analyzed according to the theoretical molecular weight (MW) and and aligned using multidimensional algorithm protein isoelectric point (pI) of the proteins identified. A repre - map software (MAProMa), which provides quantitative sentative example is reported in Additional file  3, which indexes based on the SEQUEST score, such as DAVE and shows the 2D maps corresponding to all the proteins iden- DCI [13]. DAVE is an index of the ratio between the two tified, thus confirming the possibility of characterizing conditions compared and, therefore, indicates different proteins with extreme theoretical MW values (<10  kDa) amounts of each protein under the two different condi - and proteins with extreme pI values (<4 or >10). tions. When the protein is not present in the reference In our previous study [6], the main differences at (baseline), DAVE is +2.0, whereas if the protein is not 12  months affected the cytoskeleton, mainly smooth present in the sample (in our case after treatment, T36), muscle and keratins, and ECM proteins. With respect DAVE is −2.0. Consequently, intermediate DAVE val- to ECM proteins, analysis of the biopsy specimens at ues indicate different amounts of protein under the two 36  months revealed the same trend of proteomic signa- conditions compared. DAVE values >|0.4| indicate sig- ture for patients who were Gal-3-positive before treat- nificant variations. Finally, the t test was used to evalu - ment (OR); in contrast, protein behavior was different ate significant differences in the average DAVE and fold in patients who were Gal-3-negative before treatment change values. (NOR). In the OR group, smooth muscle proteins (including Results myosins, tropomyosins, and actins) had decreased to very We evaluated the effect of long-term anti-IgE treatment low levels at 36 months (from 265 to 8 [aggregate score]; (36  months) on RBM thickness, eosinophilic and neu- average ln[T36/T0], about –4.0; DAVE, –1.89) in the OR trophilic infiltrates, and proteomic profiles by analyzing group (Gal-3-positive biopsies). In the NOR group (Gal- bronchial biopsy specimens from patients with severe 3-negative biopsies), protein levels increased from 5 to 74 asthma (Table 1). (average ln[T36/T0], around +2.5; DAVE, +1.75) (Fig. 3; Additional files 1, 4). The difference between the two Histological analysis groups was almost significant (p < 0.01). All bronchial biopsies were evaluated using morphomet- As for keratins, the trend at 36  months of treatment ric analysis of RBM thickness. was more consistent for OR (Gal-3 positive) (about six- Differences were observed in the behavior of the two fold less, from 1800 to 300 as the average score; average groups (OR and NOR) between baseline and 36  months ln[T36/T0], −2; DAVE −1.45) than NOR (both DAVE of anti-IgE treatment. and average ln[T36/T0], low stringent). In this case the Riccio et al. Clin Transl Allergy (2017) 7:6 Page 5 of 10 Bronchial reticular basement membrane (RBM) thickness changes NOR OR BASELINE AFTER 12M AFTER 36M Fig. 1 Changes in bronchial RBM thickness in omalizumab responders (OR) and non-omalizumab responders (NOR) at baseline and after 12 and 36 months of anti-IgE treatment the decreasing trend at 36 months. Specifically, periostin was reduced in all OR patients (from 117 to 0 average score; average ln[T36/T0], −4.6; DAVE −2.0), whereas in the NOR group, it was either stable (NOR3) or increased (NOR1 and NOR4), except for NOR2, in whom the periostin level decreased. Figure  5 shows the behav- ior (ln[T36/T0]) of periostin in OR and NOR over time (p < 0.05; see also Additional files 1, 4). Gal-3 was stable at 12 months, but was absent in OR at 36  months (Gal-3-positive at T0). Gal-3 was not detect- able at 36 months in NOR (Gal-3-negative at T0) (Addi- tional file 5). As reported above, NOR patients, who were char- acterized as Gal-3-negative at T0 (before treatment), presented the same behavior after 36  months of treat- Fig. 2 Evaluation of bronchial eosinophilic infiltration (cells/mm ) in ment; however, the trend for one patient (NOR2) after omalizumab responders (OR) and non-omalizumab responders (NOR) 36  months of treatment was more similar to that of OR using the natural logarithm of the score fold change, (ln[ T36/T0]). Negative values indicate a decrease in the eosinophil count at T36; than NOR patients for each of the main parameters mon- positive values indicate an increase in the eosinophil count at T36. itored. In fact, in NOR2, both smooth muscle proteins OR and NOR are classified according to whether they experienced and keratins had decreased or were absent at 36 months, a reduction in RBM thickness (OR) or no reduction in RBM thickness as was the case for OR patients. Similarly, eosinophils (NOR) after 12 months of anti-IgE treatment and periostin were, respectively, reduced and absent in NOR2, as in OR and in contrast to the remaining NOR. Table 2 summarizes the levels of the main protein classes difference was not significant (p  = 0.19; see Fig. 4; Addi- and eosinophils before treatment (T0) and 36 months after tional files 1, 4). treatment for each OR and NOR sample. Of note, both the ECM proteins, which were present almost exclusively at ln[T36/T0] (Additional file  1) and DAVE index (Additional baseline in OR (Gal-3-positive) patients, also maintained Bronchial RBM thickness (µm) Riccio et al. Clin Transl Allergy (2017) 7:6 Page 6 of 10 Clustering analysis Protein lists from the 36-month biopsy specimens were compared and used to perform an unsupervised clus- tering analysis. Thus, OR (Gal-3-positive), who were correctly segregated at T0 (see Fig.  6a), were again con- firmed to be part of the same group at 36  months. It is noteworthy that patient NOR2 was in this group. In addi- tion, the remaining three NOR patients (NOR1, NOR3, and NOR4, all Gal-3-negative) were segregated in a dif- ferent cluster (see Fig.  6b), thus confirming them to be part of the same NOR group again at 36 months of treat- ment. These findings confirm the different behavior at 36 months of patient NOR2, who was segregated with the OR group (see Table 2). Fig. 3 Changes in abundance levels, expressed as the natural loga- rithm of the score fold change (ln[ T36/T0]) for smooth muscle pro- Discussion teins (SMPs) in omalizumab responders (OR) and non-omalizumab Biomarkers of severe asthma are currently a major responders (NOR) at baseline ( T0) and after 36 months ( T36) of treat- research area [14]. Nonetheless, available biomarkers [15] ment. A negative value indicates a decrease in SMPs at T36; a positive are still far from enabling us to select appropriate candi- value indicates an increase in SMPs at T36. OR and NOR are classified dates for biologics [16, 17]. The need for molecular analy - according to whether they experienced a reduction in RBM thickness (OR) or no reduction in RBM thickness (NOR) after 12 months of anti- sis to define phenotypes has been discussed [18], and IgE treatment omics sciences have been proposed as a useful approach [19]. In this context, current advances in proteomics, mainly those based on mass spectrometry, facilitate the discov- ery-driven studies of new biomarkers in respiratory dis- eases and improve the clinical reliability of biomarkers for asthma [20]. The main focus of the present study was the extension of the histological and proteomic analyses to 36  months in a group of patients with severe asthma treated with omalizumab that we had previously studied at baseline and at 12  months [6]. In the previous study, we did not have control biopsy specimens from untreated severe asthmatic patients and/or healthy persons, since bron- choscopy and bronchial biopsy are invasive procedures that are not readily applicable in these groups. Although the number of samples was not large for the usual statistical analysis, the data obtained illustrate a Fig. 4 Changes in abundance levels, expressed as the natural loga- specific trend for both groups of patients (OR and NOR). rithm of the score fold change (ln[ T36/T0]) for keratins in omalizumab In particular, our results showed that specimens from OR responders (OR) and non-omalizumab responders (NOR) at baseline patients (Gal-3-positive at baseline) confirm the reduc - ( T0) and after 36 months ( T36). A negative value indicates a decrease tion in the thickness of bronchial RBM after 36  months in keratins at T36; a positive value indicates an increase in keratins at of treatment with anti-IgE, thus emphasizing the possi- T36. In this case, the p value is high; however, if NOR2 is excluded, the p value falls to 0.002 (Additional file 1). OR and NOR are classified ble role of omalizumab in reducing airway remodeling according to whether they experienced a reduction in RBM thickness in some cases of persistent allergic asthma [6]. As for (OR) or no reduction in RBM thickness (NOR) after 12 months of anti- molecular components, proteomic analysis showed that IgE treatment mainly ECM, keratin, and smooth muscle proteins were further reduced at 36  months of therapy in Gal-3-posi- tive biopsies. file  6) confirm the different trends for OR and NOR in the Although Gal-3 was absent in OR at 36  months, this parameters examined at baseline and T36. They also high - does not put OR in the same situation as NOR; in fact, light the different behavior of NOR2 at T36. OR presented a different proteomic profile, consistent Riccio et al. Clin Transl Allergy (2017) 7:6 Page 7 of 10 low serum periostin. This predictive role for periostin was not confirmed in recent clinical trials [22]. In the present study, periostin was not consistently present in the bron- chial biopsies of either group, and levels decreased fur- ther only in OR after 36 months of treatment. Therefore, based on our data, periostin cannot be regarded as a pre- dictive biomarker of modulation of airway remodeling. With regard to the bronchial inflammatory pat - tern, high eosinophil counts in the bronchi have been observed in patients who are eligible for anti-IgE [23] and can be considered one of the biomarkers of the Th2 asthma phenotype. Omalizumab is effective in the treat - ment of eosinophilic airway inflammation in patients with allergic asthma [24]. In the present study, a further Fig. 5 Changes in abundance levels expressed as the natural loga- decrease (compared with 12  months) in eosinophils rithm of the score fold change (ln[ T36/T0]) for periostin in omali- in the biopsy specimens was detected in patients who zumab responders (OR) and non-omalizumab responders (NOR) at were Gal-3-positive at baseline (OR), whereas eosino- baseline ( T0) and after 36 months ( T36) of anti-IgE treatment. A nega- phil counts increased in Gal-3-negative patients (NOR). tive value indicates a decrease in periostin at T36; a positive value These findings are consistent with recent data reporting indicates an increase in periostin at T36. OR and NOR are classified according to whether they experienced a reduction in RBM thickness increased and decreased Gal-3 levels in eosinophilic and (OR) or no reduction in RBM thickness (NOR) after 12 months of anti- neutrophilic asthma, respectively [25]. Gal-3 might also IgE treatment predict decreased eosinophilic inflammation in the bron - chi in patients receiving omalizumab. Finally, the further decrease in smooth muscle proteins with the reduction in RBM thickness and eosino- (i.e., actins and myosins) after 36 months is interesting and phil counts. In this context, the lack of Gal-3 in OR at thus supports the hypothesis of improved pulmonary func- 36 months, combined with the decrease in protein levels tion observed in some omalizumab-treated asthmatics [26]. reported above, may be correlated with an improvement If this was the case, Gal-3 would be a very useful tool for in remodelling and inflammatory state. predicting clinical improvement in pulmonary function. With respect to periostin, Hanania et al. [21] found that This observation is of particular clinical interest, since the patients with high serum periostin had fewer exacerba above-mentioned studies showed that, as far as FEV is tions after treatment with omalizumab than patients with Table 2 Levels of  eosinophils (cells/mm ) and  abundance (score) for  smooth muscle proteins, periostin, and  keratins at baseline (T0) and after 36 months of anti-IgE treatment (T36) OR1 OR2 OR3 OR4 NOR1 NOR2 NOR3 NOR4 Eosinophils T0 19 31 58 115 0 10 10 9 T36 3 9 4 5 11 3 20 27 Smooth muscle proteins T0 288 80 198 494 0 0 20 0 T36 10 20 0 0 168 0 30 98 Periostin T0 30 160 130 150 0 30 120 0 T36 0 0 0 0 20 0 106 20 Keratins T0 1990 1902 1630 2047 1404 1910 1892 1676 T36 574 276 100 266 2106 40 1330 1400 Galectin-3 T0 10 10 30 20 0 0 0 0 T36 0 0 0 0 0 0 0 0 Increased values (or substantially unchanged) at T36 compared to T0 are reported in italic. For more details about available proteins see Additional file 2 Riccio et al. Clin Transl Allergy (2017) 7:6 Page 8 of 10 a Baseline b After 36-months Fig. 6 Unsupervised hierarchical clustering of protein lists obtained by analyzing lung biopsies from omalizumab responders (OR, blue) and non- omalizumab responders (NOR, red). a Before (baseline). b After 36 months of anti-IgE treatment. The clusters obtained were based on SEQUEST score value and proved to be well structured (agglomerative coefficients >0.7). OR and NOR are classified according to whether they experienced a reduction in RBM thickness (OR) or no reduction in RBM thickness (NOR) after 12 months of anti-IgE treatment concerned, not all patients improved with omalizumab. In PROXIMA involves an ancillary study to explore protein responsive patients, omalizumab can be considered a dis- biomarkers in urine and plasma and to characterize them ease-modifying treatment and Gal-3 a potential biomarker. according to severe allergic asthma and treatment effects. Our findings confirm the role of Gal-3 as a biomarker In this context, the possible correlation between Gal-3 of long-term reduction in bronchial RBM, eosinophil and evolution of pulmonary function with omalizumab inflammation, and muscular components in omali - will be investigated in greater depth. zumab-treated severe asthmatics. In addition, our data support the molecular model, indicating that the pres- Conclusions ence of Gal-3 may have a role in the dissociation of the We investigated the long-term effects of anti-IgE treat - IgE–FcεRI complex and enhance the effectiveness of anti- ment. In particular, proteomic signatures combined with IgE therapy for remodelling in Gal-3-positive patients histological data enabled us to confirm the following: [7]. Further studies based on larger samples or a cellular model are required to confirm the role of Gal-3 in the • In bronchial biopsy samples from severe asthmatics, dissociation of the IgE–FcεRI complex. Gal-3 acts as a biomarker of modulation of airway Bronchial analysis is useful for investigating the origin remodeling upon treatment with omalizumab. and mechanisms of disease; however, in clinical practice, • There seems to be a relationship between Gal-3 posi - it is necessary to analyze less invasive samples, such as tivity and improvement in pulmonary function. urine and blood. We intend to perform such an analysis • In severe asthmatics treated with omalizumab, Gal-3 in the context of the PROXIMA study (Patient-Reported is a sufficiently effective biomarker of both short- and Outcomes and Xolair In the Management of Asthma), long-term reduction in the thickening of bronchial an observational, multicenter, cross-sectional, prospec- RBM, eosinophilic inflammation, and fibrotic and tive cohort study conducted at 25 centers [27]. In fact, muscular protein components. OR OR1 1 OR1 OR1 OR2 OR OR4 4 OR4 OR OR41 OR OR2 2 OR2 OR3 OR2 OR3 OR3 OR4 NOR2 NOR3 NOR3 NOR3 NOR1 NOR1 NOR1 OR3 NOR1 NOR1 NOR2 NOR2 NOR2 NOR4 NOR4 NOR4 NOR3 Riccio et al. Clin Transl Allergy (2017) 7:6 Page 9 of 10 study, data analysis and interpretation, and drafting of the manuscript. DFL Additional files performed the experiments, analyzed the data, and helped to draft the manu- script. RR performed the experiments, analyzed the data, and interpreted Additional file 1. Natural logarithm (ln) of fold change between baseline the results. DD analyzed the data and interpreted the results. BL performed ( T0) and long term anti-IgE treatment ( T36) for each subject. Significant the experiments. DRW recruited the patients and collected clinical data. MC values (ln[Fold Change]>|0.6| are are reported in red or blue: positive performed bronchoscopy and collected bronchial biopsy samples. CA contrib- (blue) and negative (red) values indicate increase and decrease at T36, uted to data analysis and participated in drafting the manuscript and in the respectively. Specifically, for ln(Fold Change) >0.6 (blue) increase at T36; critical review thereof. CGW was the lead investigator and was responsible for on the contrary, if it is <-0.6 (red) decrease at T36. SMPs: smooth muscle designing the study and drafting the manuscript. Author information: RAM proteins. Actual Fold Change was 0 (0/n, or 0/0, see Table 2); for avoiding and MP contributed equally to this work. All authors read and approved the invalid logarithm it was put at 1/n or 1/1; Actual Fold Change ∞ (n/0); for final manuscript. avoiding invalid logarithm it was put at n/1. * p-value; T-test2 is without NOR2 subject, because at T36 its behaviour is similar to ORs (se Fig. 6); in Author details bold significant T-tests; ** #/mm2 ratio; *** score ratio. Respiratory Diseases and Allergy Unit, IRCCS AOU San Martino-IST, Univer- sity of Genoa, Genoa, Italy. Proteomics and Metabolomics Unit, Institute Additional file 2. Complete list of identified proteins in OR (Responders) for Biomedical Technologies, CNR, Milan, Italy. National Centre for Respira- and NOR (Non-Responders) bronchial biopsies collected after 36 months tory Pharmacoeconomics and Pharmacoepidemiology, CESFAR, Verona, Italy. of anti-IgE treatment. Accession number NCBI, Uniprot entry, Protein 4 5 Respiratory Unit, Mater Salutis Hospital, Legnago, Verona, Italy. Department name, pI and MW, Spectral Count (SpC) and Score were reported for each of Biomedical Sciences, Personalized Medicine Clinic Asthma and Allergy, protein. The symbol (*) was used to indicate the smooth muscle proteins. Humanitas University, Rozzano, Milan, Italy. The symbol (†) was used to indicate the periostin. The symbol (°) was used to indicate the keratins. Acknowledgements Additional file 3. Virtual 2D map in logarithmic scale of FFPE bronchial This work was partially supported by ARMIA (Associazione Ricerca Malattie biopsies, generated using MAProMa software (546 proteins). Proteins are Immunologiche ed Allergiche) and the Italian CNR FLAGSHIPs Projects, funded plotted according to their theoretical pI and MW. A color/shape code by the MIUR, “InterOmics”. The authors thank Danila Vella for statistical support. is assigned to each protein according to SEQUEST score value. Proteins with score ≤15 are reported as yellow/triangle, proteins with score ≥35 Competing interests are reported as red/circle, and proteins in the range 15–35 are reported as The authors declare that they have no competing interests. blue/square. The protein lists are reported in Additional file 2. Availability of data and materials Additional file 4. Changes of abundance levels, calculated by DAVE The datasets supporting the conclusions of this article are included within the algorithm from MAPROMA software [9], for eosinoplis, smooth muscle article and its additional files. proteins, periostin, keratins and Gal-3 in OR and NOR patients at baseline ( T0) and after 36 months ( T36) of anti-IgE treatment. Negative value indi- Ethics approval cates decrease at T36; on the contrary, positive value indicates increase at This study was approved by the local Ethics Committee. T36. OR and NOR classification is related to reduction (OR) or not (NOR) of RBM thickness after 12 months of anti-IgE treatment. Funding Additional file 5. Changes of abundance levels, expressed as natural The authors declare that the research was not funded. logarithm of score fold change (ln[ T36/T0]), for Gal-3 in OR and NOR patients at baseline ( T0) and after 36 months ( T36) of anti-IgE treatment. Received: 13 September 2016 Accepted: 22 February 2017 Negative value indicates decrease of Gal-3 at T36; on the contrary, positive value indicates increase of Gal-3 at T36 (see Additional file 1). OR and NOR classification is related to reduction (OR) or not (NOR) of RBM thickness after 12 months of anti-IgE treatment. Additional file 6. Differential Analysis (DAVE index) between baseline References ( T0) and long term anti-IgE treatment ( T36) for each subject, using DAVE 1. Brasier AR, Ju H. Analysis and predictive modeling of asthma phenotypes. index from MAPROMA software [9]. Significant values (DAVE > |0.4|) are Adv Exp Med Biol. 2014;795:273–88. reported in red or blue: positive (blue) and negative (red) values indicate 2. Wenzel SE. Asthma phenotypes: the evolution from clinical to molecular increase and decrease at T36, respectively. SMPs: smooth muscle proteins. approaches. Nat Med. 2012;18:716–25. Of note, DAVE algorithm, tpical of proteomics evaluation, was also 3. Jarjour NN, Erzurum SC, Bleecker ER, Calhoun WJ, Castro M, Comhair SA, applied to eosinophils; the obtained values resulted in good agreement et al. Severe asthma: lessons learned from the National Heart, Lung, and with evaluation obtained by ln[ T36/T0] (see also Additional file 1 and Blood Institute Severe Asthma Research Program. Am J Respir Crit Care compare Fig. 2 and Additional file 4). * p-value; T-test2 is without NOR2 Med. 2012;185:356–62. subject, because at T36 its behaviour is similar to ORs (see Fig. 6); in bold 4. Gandhi NA, Bennett BL, Graham NM, Pirozzi G, Stahl N, Yancopoulos GD. significant T-tests. Targeting key proximal drivers of type 2 inflammation in disease. Nat Rev Drug Discov. 2016;15(1):35–50. 5. Fajt ML, Wenzel SE. Asthma phenotypes and the use of biologic medica- tions in asthma and allergic disease: the next steps toward personalized Abbreviations care. J Allergy Clin Immunol. 2015;135:299–310. OR: omalizumab responder; NOR: non-omalizumab responder; IgE: immuno- 6. Riccio AM, Dal Negro RW, Micheletto C, De Ferrari L, Folli C, Chiappori A, globulin E; MudPIT: multidimensional proteomic identification technology; et al. Omalizumab modulates bronchial reticular basement membrane Gal-3: galectin-3; ECM: extracellular matrix; Th2: T-helper type-2 lymphocytes; thickness and eosinophil infiltration in severe persistent allergic asthma FeNO: fractional exhaled nitric oxide; IL: interleukin; pI: isoelectric point; RBM: patients. Int J Immunopathol Pharmacol. 2012;25:475–84. reticular basement membrane; FEV : forced expiratory volume in the first 7. Mauri P, Riccio AM, Rossi R, Di Silvestre D, Benazzi L, De Ferrari L, et al. second; MW: molecular weight; ln: natural logarithm; FC: fold change (ratio Proteomics of bronchial biopsies: galectin-3 as a predictive biomarker of T36/T0); P: p value. airway remodelling modulation in omalizumab-treated severe asthma patients. Immunol Lett. 2014;162(1 Pt A):2–10. Authors’ contributions 8. Hsia CC, Hyde DM, Ochs M, Weibel ER. ATS/ERS Joint Task Force on Quan- RAM contributed to the design of the study, data acquisition, the statistical titative Assessment of Lung Structure. An official research policy state - analysis, and drafting of the manuscript. MP contributed to design of the ment of the American Thoracic Society/European Respiratory Society: Riccio et al. Clin Transl Allergy (2017) 7:6 Page 10 of 10 standards for quantitative assessment of lung structure. Am J Respir Crit 20. Rossi R, De Palma A, Benazzi L, Riccio AM, Canonica GW, Mauri P. Care Med. 2010;181:394–418. Biomarker discovery in asthma and COPD by proteomic approaches. 9. Chetta A, Zanini A, Foresi A, Del Donno M, Castagnaro A, D’Ippolito R, Proteomics Clin Appl. 2014;8:901–15. et al. Vascular component of airway remodeling in asthma is reduced by 21. Hanania NA, Wenzel S, Rosen K, Hsieh HJ, Mosesova S, Choy DF, et al. high dose of fluticasone. Am J Respir Crit Care Med. 2003;167:751–7. Exploring the effects of omalizumab in allergic asthma: an analysis of bio - 10. Mauri P, Scigelova M. Multidimensional protein identification technology markers in the EXTRA study. Am J Respir Crit Care Med. 2013;187:804–11. for clinical proteomic analysis. Clin Chem Lab Med. 2009;47:636–46. 22. Hanania NA, Korenblat P, Chapman KR, Bateman ED, Kopecky P, Paggiaro 11. Zhao Y, Karypis G. Data clustering in life sciences. Mol Biotechnol. P, et al. Efficacy and safety of lebrikizumab in patients with uncontrolled 2005;31:55–80. asthma (LAVOLTA I and LAVOLTA II): replicate, phase 3, randomised, dou- 12. Vigani G, Di Silvestre D, Agresta AM, Donnini S, Mauri P, Gehl C et al. ble-blind, placebo-controlled trials. Lancet Respir Med. 2016;4:781–96. Molybdenum and iron mutually impact their homeostasis in cucumber 23. Busse W, Spector S, Rosén K, Wang Y, Alpan O. High eosinophil count: (Cucumis sativus) plants. New Phytol. 2017;213(3):1222–41. a potential biomarker for assessing successful omalizumab treatment 13. Mauri P, Dehò G. A proteomic approach to the analysis of RNA degrado- effects. J Allergy Clin Immunol. 2013;132:485–6. some composition in Escherichia coli. Methods Enzymol. 2008;447:99–117. 24. Djukanović R, Wilson SJ, Kraft M, Jarjour NN, Steel M, Chung KF. Eec ff ts 14. Arron JR, Choy DF, Scheerens H, Matthews JG. Noninvasive biomarkers of treatment with anti-immunoglobulin E antibody omalizumab on that predict treatment benefit from biologic therapies in asthma. Ann airway inflammation in allergic asthma. Am J Respir Crit Care Med. Am Thorac Soc. 2013;10(Suppl):S206–13. 2004;170:583–93. 15. Leung TF, Ko FW, Wong GW. Recent advances in asthma biomarker 25. Gao P, Gibson PG, Baines KJ, Yang IA, Upham JW, Reynolds PN, Hodge S, research. Ther Adv Respir Dis. 2013;7:297–308. James AL, Jenkins C, Peters MJ, Zhang J, Simpson JL. Anti-inflammatory 16. Chung KF. New treatments for severe treatment-resistant asthma: target- deficiencies in neutrophilic asthma: reduced galectin-3 and IL-1RA/IL-1β. ing the right patient. Lancet Respir Med. 2013;1:639–52. Respir Res. 2015;24(16):5. 17. De Ferrari L, Chiappori A, Bagnasco D, Riccio AM, Passalacqua G, Canonica 26. Baena-Cagnani CE, Teijeiro A, Canonica GW. Four-year follow-up in GW. Molecular phenotyping and biomarker development: Are we on our children with moderate/severe uncontrolled asthma after withdrawal way towards targeted therapy for severe asthma? Expert Rev Respir Med. of a 1-year omalizumab treatment. Curr Opin Allergy Clin Immunol. 2016;10:29–38. 2015;15:267–71. 18. Wenzel SE. Complex phenotypes in asthma: current definitions. Pulm 27. Chiappori A, De Ferrari L, Folli C, Mauri P, Riccio AM, Canonica GW. Pharmacol Ther. 2013;26:710–5. Biomarkers and severe asthma: a critical appraisal. Clin Mol Allergy. 19. Holgate ST. Trials and tribulations in identifying new biologic treatments 2015;13:20. for asthma. Trends Immunol. 2012;33:238–46. Submit your next manuscript to BioMed Central and we will help you at every step: • We accept pre-submission inquiries • Our selector tool helps you to find the most relevant journal • We provide round the clock customer support • Convenient online submission • Thorough peer review • Inclusion in PubMed and all major indexing services • Maximum visibility for your research Submit your manuscript at www.biomedcentral.com/submit

Journal

Clinical and Translational AllergySpringer Journals

Published: Mar 9, 2017

There are no references for this article.