Phenotypic plasticity and targeting of Siglec-F high CD11c low eosinophils to the airway in a murine model of asthma

Allergy BRIEF COMMUNICATION Phenotypic plasticity and targeting of Siglec-F high CD11c low eosinophils to the airway in a murine model of asthma H. Abdala Valencia 1, L. F. Loffredo 1, A. V. Misharin 2 & S. Berdnikovs 1 1 Division of Allergy-Immunology, Northwestern University Feinberg School of Medicine; 2 Division of Pulmonary and Critical Care, Northwestern University Feinberg School of Medicine, Chicago, IL, USA To cite this article: Abdala Valencia H, Loffredo LF, Misharin AV, Berdnikovs S. Phenotypic plasticity and targeting of Siglec-F high CD11c low eosinophils to the airway in a murine model of asthma. Allergy 2016; 71: 267 271. Keywords allergic inflammation; eosinophils; integrins; phenotypic plasticity; recruitment kinetics. Correspondence Sergejs Berdnikovs, Division of Allergy- Immunology, Northwestern University Feinberg School of Medicine, McGaw Building, Room M-302, 240 East Huron Street, Chicago, IL 60611, USA. Tel.: 312-503-6924 Fax: 312-503-0078 E-mail: s-berdnikovs@northwestern.edu Accepted for publication 22 September 2015 DOI:10.1111/all.12776 Edited by: Michael Wechsler Abstract Eosinophil recruitment in asthma is a multistep process, involving both trans-endothelial migration to the lung interstitium and trans-epithelial migration into the airways. While the trans-endothelial step is well studied, trans-epithelial recruitment is less understood. To contrast eosinophil recruitment between these two compartments, we employed a murine kinetics model of asthma. Eosinophils were phenotyped by multicolor flow cytometry in digested lung tissue and bronchoalveolar lavage (BAL) simultaneously, 6 h after each ovalbumin (OVA) challenge. There was an early expansion of tissue eosinophils after OVA challenge followed by eosinophil buildup in both compartments and a shift in phenotype over the course of the asthma model. Gradual transition from a Siglec- F med CD11c to a Siglec-F high CD11c low phenotype in lung tissue was associated with eosinophil recruitment to the airways, as all BAL eosinophils were of the latter phenotype. Secondary microarray analysis of tissue-activated eosinophils demonstrated upregulation of specific integrin and chemokine receptor signature suggesting interaction with the mucosa. Using adhesion assays, we demonstrated that integrin CD11c mediated adhesion of eosinophils to fibrinogen, a significant component of epithelial barrier repair and remodeling. To the best of our knowledge, this is the only report to date dissecting compartmentalization of eosinophil recruitment as it unfolds during allergic inflammation. By capturing the kinetics of eosinophil phenotypic change in both tissue and BAL using flow cytometry and sorting, we were able to demonstrate a previously undocumented association between phenotypic shift of tissue-recruited eosinophils and their trans-epithelial movement, which implicates the existence of a specific mechanism targeting these cells to mucosal airways. Recruitment of eosinophils to allergen-induced lung tissue is a hallmark of allergic airway inflammation. Eosinophil recruitment is a multistep process involving both trans-endothelial migration from peripheral circulation into the lung interstitial space and trans-epithelial migration from the interstitium into the lung bronchoalveolar (mucosal) space. While trans-endothelial migration of eosinophils is well studied, dynamics of trans-epithelial recruitment are less understood. Murine models of ovalbumin (OVA)/alum-induced allergic inflammation remain an essential tool for approximating leukocyte recruitment in human asthma. In these models, eosinophils are commonly quantified in cytospin preparations of bronchoalveolar lavage (BAL). However, this compartment represents only an endpoint in eosinophil migration and does not allow for a level of resolution necessary to address the sequence of events leading to priming and compartmentalization of eosinophil recruitment. To remedy this, we monitored eosinophil recruitment simultaneously in digested lung tissue and BAL after each allergen challenge, using multicolor flow cytometry. This is the first study to dissect compartmentalization of eosinophil recruitment, tracking eosinophil phenotypes as they progressively populate and move across lung compartments during sequential challenges. Doing so has allowed us to associate gradual acquisition of a Siglec-F high CD11c low phenotype with movement of eosinophils from the lung interstitium into the airways. Integrin profiling of lung-recruited eosinophils further suggests that eosinophil phenotypic plasticity during allergic inflammation Allergy 71 (2016) 267 271 2015 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd 267

Trans-epithelial recruitment of eosinophils Abdala Valencia et al. is consistent with their multistep recruitment and specific movement into the airway. Materials and methods Kinetics model of allergic lung inflammation All experiments were performed with 6- to 12-week-old female BALBc/J mice (Jackson Labs, Bar Harbor, ME, USA). The mice were sensitized by intraperitoneal injection (200 ll) of OVA grade V 10 lg/alum or saline/alum on days 0 and 7 and then challenged intranasally on days 15 (Ch1), 17 (Ch2), 19 (Ch3), 21 (Ch4), 23 (Ch5) and 25 (Ch6) with OVA grade VI (50 lg/50 ll saline) or saline alone. Tissues were harvested and processed for analysis at 6 h after each OVA challenge. Bronchoalveolar lavage, lung digestion, flow cytometry, and eosinophil sorting Bronchoalveolar lavage was performed by gently lavaging lungs with ice-cold 19 PBS through the cannulated trachea. Following the lavage, lungs were removed from the mediastinal cavity and lobes of the left lung were digested in 1 mg/ml collagenase D and 0.2 mg/ml DNAse I (Roche, Indianapolis, IN, USA) in preparation for flow cytometry. Digested tissue was filtered through sterile mesh and incubated in 19 BD PharmLyse Lysing Buffer (BD Biosciences, San Jose, CA, USA) to lyse red blood cells. Preparations were stained for live/dead exclusion with Aqua dye (Molecular Probes) followed by incubation with CD16/CD32 FC Block (BD Pharmingen, San Jose, CA, USA). Antibody cocktail was added directly to blocked samples and incubated for 30 min at 4 C. Samples were acquired on a BD LSRII flow cytometer (BD Biosciences). All centrifugation steps were carried out at 300 g for 5 min. BAL cells were pelleted by centrifugation, washed, and prepared as described above, starting with the live/dead step. We used the following panels of fluorescent dyes and fluorochrome-conjugated antibodies: (i) Live/Dead Aqua fluorescent dye (Molecular Probes); (ii) CD45 (clone 30-F11/Biolegend/AF700 and FITC); (iii) CD11b (clone M1/70/BD Biosciences/PE-CF594 and APC/ Cy7); (iv) Siglec-F (clone E50-2440/BD Biosciences/PE and APC); (v) CD11c (clone N418/Biolegend/PE/Cy7); (vi) Ly6G (clone 1A8/Biolegend/PerCP/Cy5.5 and AF700); (vii) Ly6C (clone AL-21/eBioscience/eFluor450); and (viii) CD64 (clone X54-5/7.1/Biolegend/PE). Bead compensation (OneComp; ebioscience, San Diego, CA, USA, and ArC; Molecular Probes beads), gating and data analysis were performed using FlowJo v.10 (TreeStar, Inc., Ashland, OR, USA). Only live, single, hematopoietic (CD45 + ) cells were used in all analyses. Fluorescence Minus One (FMO) controls were used to set up gate boundaries. Eosinophil cell sorting for cytospins was performed on a FACSAria III (BD Biosciences) instrument using a 100- lm nozzle. Leukocyte populations were identified as follows: (i) Eosinophils: CD11b + CD64 Ly6G low CD11c /low Siglec-F med/ high ; (ii) Alveolar macrophages: CD11b CD64 + Ly6G CD 11c high Siglec-F high ; and (iii) Neutrophils: CD11b + CD64 Ly6- G high CD11c Siglec-F. The complete gating strategy is illustrated in Fig. 1A and Fig. S1. Adhesion assays Fibrinogen (Oxford Biomedical Research, Oxford, MI, USA) and VCAM-1 (BD Biosciences) were plated in 96-well plates in TBS ph 8.0 (50 lg/ml) and incubated overnight at 4 C. Wells were blocked with neat FBS for 1 h at 37 C. Eosinophils derived from spleens of NJ1638 mice (kind gift of Dr. James Lee, Mayo Clinic Arizona) were labeled with 2 lm CFSE (Biolegend, San Diego, CA, USA) and blocked with 10 lg/ml rat anti-cd11b (Biolegend), Armenian Hamster anti-cd11c (Biolegend) or corresponding normal IgG controls (Santa Cruz Biotechnology, Dallas, TX, USA). Treated or untreated cells (10 5 cells in 100 ll) were added to wells for 60 min at 37 C. Additional eosinophils were plated in noncoated wells and serially diluted to generate a standard curve. The plate was immediately read on a Gemini EM Microplate Reader (Molecular Devices, Sunnyvale, CA, USA) at an excitation of 485/20 and emission of 528/20. The CFSE fluorescence levels were used as a measure of eosinophil adherence by examining the average fluorescence of the samples relative to the standard curve. Secondary microarray data analysis and statistics We performed secondary data analysis of a microarray study comparing eosinophils sorted from lung digests of saline vs OVA-challenged wild-type mice (study GSE57757, GEO, NCBI). Raw datasets were analyzed using the R Bioconductor package, implementing a moderated t-test with FDR correction. Hierarchical clustering was done on differentially expressed genes with a significance cut-off of 0.05 using average linkage algorithm (Cluster/TreeView). We used ANOVA and Dunnett s post hoc analysis to test for significant differences in kinetic responses of leukocytes (SYSTAT 10, Systat Software, Inc., San Jose, CA, USA). Results and discussion Lung resident eosinophils Although eosinophils are not usually perceived as resident in homeostatic lung tissue, we reproducibly detected and sorted a small population of eosinophils in lung tissue homogenates of na ıve adult mice, which averaged 1.5% of total CD45 + hematopoietic cells (Figs 1 and 2, and Fig. S2). No eosinophils were detected in the na ıve mouse BAL. All lung eosinophils found in na ıve mice exhibited a Siglec-F med CD11c phenotype, consistent with existing strategies differentiating lung eosinophils and alveolar macrophages (Fig. 2A) (1). Typically, resident eosinophils are found in the mucosa of the intestine, uterus, and thymus, where they likely have homeostatic functions (2). Interestingly, resident eosinophil populations in these organs express CD11c (3, 4). Eosinophil recruitment The first OVA challenge doubled the eosinophil population in the lung tissue (3%), although no eosinophils were detected in the BAL at this time. The tissue population stea- 268 Allergy 71 (2016) 267 271 2015 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd

Abdala Valencia et al. Trans-epithelial recruitment of eosinophils Figure 1 Flow cytometry analysis of kinetics of leukocyte recruitment in a murine model of allergic inflammation. (A) Basic gating strategy used to identify three leukocyte populations of primary interest: eosinophils, neutrophils, and alveolar macrophages. (B) Total counts of alveolar macrophages, neutrophils, and eosinophils in different lung compartments by flow cytometry during allergic inflammation. Left, lung tissue homogenates; right, bronchoalveolar lavage (matching animals). n.s., no statistical significance; red asterisks represent significant change (P < 0.05) in number of eosinophils compared to saline controls; green asterisks represent significant change in number of neutrophils compared to saline controls; black lines and asterisks represent statistically significant differences between neutrophils and eosinophils at any given challenge (P < 0.05). Shown is one representative kinetics profile of two independent experiments (n = 2 4 replicates per challenge per experiment). dily expanded with each consecutive challenge, constituting 40% of CD45 + cells by challenge 6 (Fig. 1B and Fig. S2). In the BAL, eosinophils became detectable only after the second challenge. Numbers of eosinophils in the BAL were not greater than total numbers of eosinophils in the lung. However, proportional representation of eosinophils relative to other cell types was higher in the BAL than in the lung, constituting up to 75% of all CD45 + cells in the BAL during challenges 4 6, which suggests that eosinophils may have a specific function in airway lumens. Phenotypic plasticity of lung eosinophils We identified two distinct populations of eosinophils in the lung tissue of OVA-challenged mice: Siglec-F med CD11c and Siglec-F high CD11c low (Fig. 2). Upregulation of Siglec-F and CD11c by tissue eosinophils became more pronounced with more challenges, which synchronized with increase in eosinophils in the BAL. BAL-recruited eosinophils only exhibited one eosinophil phenotype (Siglec-F high CD11c low ), suggesting that only this population crosses the epithelial barrier and enters the airways. We sorted all populations and confirmed by cytospin that they contained pure eosinophils (Fig. 2B). Interestingly, morphology of eosinophils changed from conventional ring-shaped nuclei in na ıve mice to more segmented nuclei in Siglec-F med CD11c eosinophils sorted from lungs of OVA-challenged mice, to highly segmented nuclei and vacuole-containing cytoplasm in CD11c + eosinophils found in lungs and BAL of challenged mice (Fig. 2B). CD11c upregulation by activated murine lung eosinophils has also been reported during allergen exposure (5, 6), helminth infection (7), in IL-5 transgenic mice (6), and papaininduced airway damage (8), although no previous studies contextualized this information with regard to eosinophil recruitment. Based on our results, we propose that the Siglec-F high CD11c low phenotype switch is associated with movement of eosinophils into the mucosal compartment. In support of this, CD11c is a mucosal-docking integrin, usually linked to cell types associated with the mucosa: alveolar macrophages, plasmacytoid dendritic cells and NK cells. Allergy 71 (2016) 267 271 2015 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd 269

Trans-epithelial recruitment of eosinophils Abdala Valencia et al. A B C D E Figure 2 Shift in eosinophil phenotype during allergic inflammation. (A) Phenotyping of eosinophils by flow cytometry during the course of allergic inflammation identified two populations: left gate, Siglec- F med CD11c and right gate, Siglec-F high CD11c low. (B) Cytospin preparations of different eosinophil populations sorted from na ıve and ovalbumin (OVA)-challenged lungs (209). Representative morphology of each eosinophil phenotype is shown in circles on each image (609, oil). (C) Quantitation of eosinophil phenotypes shown in A by flow cytometry: left, proportional increase in Siglec-F high population relative to total eosinophils; right, % of eosinophils expressing CD11c in Siglec-F med and Siglec-F high eosinophil populations. *All groups represented by white bars were statistically significantly different from the ones represented by black bars at any given challenge, P < 0.05. (D) Secondary microarray analysis (heatmap representation) of chemokine receptor and integrin expression by eosinophils sorted from lung tissue of saline vs OVA-challenged mice. Siglec-F (Sigec5) and CD11c (Itgax) are highlighted in red. The black box highlights upregulated integrin/chemokine receptor signature with the highest clustering on the heatmap. (E) Eosinophil adhesion to fibrinogen. FBS used as a negative adhesion control, VCAM-1 as a positive adhesion control. *P < 0.05, data shown are experimental means (n = 2 4 independent adhesion experiments). 270 Allergy 71 (2016) 267 271 2015 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd

Abdala Valencia et al. Trans-epithelial recruitment of eosinophils Additional insights into phenotypic change of lung eosinophils To gain additional clues to the role of phenotypic changes of tissue-recruited eosinophils, we performed secondary analysis of a microarray study contrasting eosinophils isolated from lung tissue of saline vs. OVA-challenged mice (three challenges; Fig. 2D) (9). Consistent with our findings, we identified a significant increase in expression of CD11c (Itgax) and Siglec-F (Siglec5) by allergen-activated eosinophils. Concurrently, these eosinophils upregulated specific chemokine receptors Ccrl2, Ccr2, and Ccr5, and integrins Itgae, Itgb1, and Itgb5, suggesting enhanced capacity for interaction with the mucosa. Trans-endothelial step-associated integrin Itga4 was downregulated, which could assist eosinophils in their release from the tissue into lumens (10). Using adhesion assays, we demonstrated that CD11c mediates adhesion of eosinophils to fibrinogen, which is a significant component of epithelial barrier repair/remodeling (Fig. 2E) (11). Upregulation of Siglec-F may represent a negative feedback mechanism, as this receptor has been shown to limit eosinophil recruitment (12, 13). The functional significance of eosinophil phenotypic transition during trans-epithelial migration and microarray phenotyping of eosinophil subsets is the subject of further investigation by our group. Acknowledgments We would like to thank Drs Bruce Bochner and Joan Cook- Mills (Northwestern University Feinberg School of Medicine) for sharing both technical and intellectual expertise in fulfillment of this project. Conflicts of interest The authors declare that they have no conflicts of interest. Author contributions Sergejs Berdnikovs and Hiam Abdala Valencia participated in conception and design of research; Sergejs Berdnikovs and Hiam Abdala Valencia performed experiments; Alexander V. Misharin sorted eosinophils; Sergejs Berdnikovs and Lucas F. Loffredo analyzed data; Sergejs Berdnikovs, Hiam Abdala Valencia, and Lucas F. Loffredo interpreted results of experiments; Lucas F. Loffredo and Sergejs Berdnikovs prepared figures; Sergejs Berdnikovs, Lucas F. Loffredo, and Hiam Abdala Valencia drafted and revised manuscript; Sergejs Berdnikovs approved final version of manuscript. Supporting Information Additional Supporting Information may be found in the online version of this article: Figure S1. Gating strategy for FACS sorting of eosinophils. Figure S2. Flow cytometry analysis of kinetics of leukocyte recruitment in a murine model of allergic inflammation (percentage data). References 1. Dyer KD, Garcia-Crespo KE, Killoran KE, Rosenberg HF. Antigen profiles for the quantitative assessment of eosinophils in mouse tissues by flow cytometry. J Immunol Methods 2011;369:91 97. 2. Lee JJ, Jacobsen EA, McGarry MP, Schleimer RP, Lee NA. Eosinophils in health and disease: the LIAR hypothesis. Clin Exp Allergy 2010;40:563 575. 3. Throsby M, Herbelin A, Pleau JM, Dardenne M. CD11c+ eosinophils in the murine thymus: developmental regulation and recruitment upon MHC class I-restricted thymocyte deletion. J Immunol 2000;165:1965 1975. 4. Carlens J, Wahl B, Ballmaier M, Bulfone-Paus S, Forster R, Pabst O. Common gamma-chaindependent signals confer selective survival of eosinophils in the murine small intestine. J Immunol 2009;183:5600 5607. 5. Rose CE Jr, Lannigan JA, Kim P, Lee JJ, Fu SM, Sung SS. Murine lung eosinophil activation and chemokine production in allergic airway inflammation. Cell Mol Immunol 2010;7:361 374. 6. Wang HB, Ghiran I, Matthaei K, Weller PF. Airway eosinophils: allergic inflammation recruited professional antigen-presenting cells. J Immunol 2007;179:7585 7592. 7. Voehringer D, Shinkai K, Locksley RM. Type 2 immunity reflects orchestrated recruitment of cells committed to IL-4 production. Immunity 2004;20:267 277. 8. Motomura Y, Morita H, Moro K, Nakae S, Artis D, Endo TA et al. Basophilderived interleukin-4 controls the function of natural helper cells, a member of ILC2s, in lung inflammation. Immunity 2014;40:758 771. 9. Wen T, Mingler MK, Wahl B, Khorki ME, Pabst O, Zimmermann N et al. Carbonic anhydrase IV is expressed on IL-5-activated murine eosinophils. J Immunol 2014;192:5481 5489. 10. Catalli AE, Thomson JV, Babirad IM, Duong M, Doyle TM, Howie KJ et al. Modulation of beta1-integrins on hemopoietic progenitor cells after allergen challenge in asthmatic subjects. J Allergy Clin Immunol 2008;122:803 810. 11. Perrio MJ, Ewen D, Trevethick MA, Salmon GP, Shute JK. Fibrin formation by wounded bronchial epithelial cell layers in vitro is essential for normal epithelial repair and independent of plasma proteins. Clin Exp Allergy 2007;37:1688 1700. 12. Zimmermann N, McBride ML, Yamada Y, Hudson SA, Jones C, Cromie KD et al. Siglec-F antibody administration to mice selectively reduces blood and tissue eosinophils. Allergy 2008;63:1156 1163. 13. Zhang M, Angata T, Cho JY, Miller M, Broide DH, Varki A. Defining the in vivo function of Siglec-F, a CD33-related Siglec expressed on mouse eosinophils. Blood 2007;109:4280 4287. Allergy 71 (2016) 267 271 2015 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd 271